Journal of Geotechnical and Geoenvironmental Engineering

Multiple barriers with basal clearance are usually adopted in engineering practice to mitigate large volumes of debris flows. This approach allows for sufficient discharge and minimizes the need for cleaning and maintenance over the service life of the barriers. Despite the effectiveness, the design of basal clearance is still based on empirical methods, which can lead to uncertainty in safety. In this study, a new dimensionless overflow number $N_{of}$ is proposed for designing dual rigid barriers with basal clearance and validated with physical and numerical results. The physical tests were conducted using a 28 m long flume model. A finite element model calibrated using the physical test results is utilized for numerical parametric study to investigate the effects of basal clearance on debris flow impact force against dual barriers. Physical and numerical results reveal that $N_{of}$ can serve to determine whether the impact force on the second barrier is governed by basal discharge or overflow from the first barrier. A threshold value of $N_{of}=1$ can be used to design dual barriers with basal clearance to achieve the lowest impact force on the second barrier. The proposed $N_{of}$ also enables engineers to predict the impact force on the second barrier by only considering the impact force from basal discharge.

Abstract

In the random field model’s consideration of the spatial variability of soil, soil properties at different locations play different roles in the reliability analysis of the foundation. Investigating the importance distribution of the random field through reliability sensitivity analysis (RSA) is beneficial for understanding how the random field affects the reliability of the foundation. However, many existing RSA methods for the random field model are deficient in terms of efficiency, accuracy, and applicability under complex engineering conditions. Consequently, this study proposes an efficient RSA method for the random field model based on the Karhunen–Loève (KL) expansion method and the first-order reliability method (FORM) to identify the important random field domain in foundation engineering. In the proposed method, the mean reliability sensitivity index (MRSI) is extended to a random field model of continuous form to characterize the importance distribution of the random field. The MRSI is analytically derived based on the results of the KL expansion method and the FORM without additional limit state function (LSF) calculations. Subsequently, the important random field domain, in which the variation of the mean of the soil property contributes significantly to the reliability index, is identified based on the MRSI. Last, two foundation engineering examples that consider the cross-correlated random fields of cohesion and friction angle, including strip footing on single-layer soil and pile in multiple-layer soil, were used to verify the proposed method. The results showed that an important random field domain with a small area dominates the variation of the reliability index of a foundation, and important random field domain area increases with autocorrelation distance (ACD). This innovative identification method holds great engineering significance, because it allows geotechnical practitioners to gain a comprehensive understanding of the failure modes and foundation treatment areas of foundations in spatially varying soil.

Practical Applications

In the random field model’s consideration of the spatial variability of soil, soil properties at different locations play different roles in the reliability analysis of the foundation. Investigating the importance distribution of the random field through RSA is beneficial for understanding how the random field affects the reliability of the foundation. However, many existing RSA methods for the random field model are deficient in terms of efficiency, accuracy, and applicability under complex engineering conditions. Consequently, this study proposes an efficient RSA method for the random field model to identify the important random field domain, in which the variation of the mean of the soil property contributes significantly to the reliability index. Two foundation engineering examples that consider the cross-correlated random fields of cohesion and friction angle, including strip footing on single-layer soil and pile in multiple-layer soil, were used to verify the proposed method. The results showed that the innovative identification will allow geotechnical practitioners to gain a comprehensive understanding of the failure modes and foundation treatment areas of foundations in spatially varying soil.

Soil properties can vary significantly within a given site, making geotechnical analysis highly amenable to a probabilistic treatment in which properties are described statistically. The random finite-element method (RFEM) is the only method available to properly account for spatial variability of soil properties and has been applied to numerous geotechnical applications. Of all the commonly encountered soil properties, permeability is generally considered to be one of the most variable. The current paper revisits steady seepage beneath a sheet pile wall of varying length in spatially random soil of finite depth, with a focus on the exit gradient. The spatial correlation length of permeability is shown to have a striking influence on the statistics of the exit gradient, with a pronounced worst-case spatial correlation length corresponding to the highest probability of downstream piping and erosion.

Abstract

The principle of detecting seepage anomalies in water-retaining earthen structures through temperature monitoring has yet to gain extensive acceptance in modern field applications. As such, this study demonstrated the practical detection of seepage anomaly from fiber optic distributed temperature sensing (DTS) through physical and numerical modeling. A high-resolution DTS, capturing temperature at a sub-cm interval, was embedded in a laboratory-scale earth dam featuring normal and anomalous seepage conditions (induced by artificial defect). Then, a fluctuating reservoir thermal load was applied, and the internal temperature of the dam was monitored. Evaluation of the temperature lag time showed a quicker detection of temperature changes on the DTS sensing locations traversed by the anomalous seepage flow. The disparity in temperature lag times across the DTS increased significantly when a severe seepage anomaly was induced by systematically channeling the reservoir water directly into the artificial defect through a small tube tapping into the reservoir. Further insight into the heat-seepage interaction was achieved through coupled hydrothermal numerical modeling using COMSOL Multiphysics.

Practical Applications

Numerous dam and levee failures worldwide linked to poor conditions and erosion have caused much damage and loss of life. This study practically demonstrated temperature analysis as a robust indicator for identifying potential issues related to seepage in earthen water infrastructures, such as dams and levees. This methodology facilitates efficient early detection, allowing for proactive intervention to avert potential failures and improve public safety. Moreover, the research findings are expected to lay the groundwork for a proof of concept, cultivating an improved understanding and appreciation of the technique among practicing engineers. Building on the success of this study, future laboratory investigations are planned to explore deeper into seepage-heat interactions and optimize the technique for enhanced practical applications, including determining factors such as the ideal location and embedment depth for the distributed temperature sensing system.

During the last decades, liquefaction damages induced by earthquakes have underlined the importance of identifying effective soil improvement techniques for mitigation purposes. Vibratory methods, such as rammed aggregate piers, are commonly used to densify sands and silty sands, erroneously neglecting the influence of the lateral stress. This paper presents the results of a series of liquefaction mitigation case studies carried out using rammed aggregate piers in Christchurch (New Zealand), Boca de Briceño (Ecuador), and Bondeno (Italy) following the 2010–2011 Canterbury seismic sequence, the 2016 Muisne earthquake, and the 2012 Emilia seismic sequence, respectively. The availability of coupled piezocone and seismic dilatometer tests before and after treatment enabled a geotechnical characterization of the three sandy sites to be made, along with estimating the at-rest lateral earth pressure coefficient, and comparing the effectiveness of the treatment at the trial sites. Finally, the paper proposes an updated procedure for liquefaction assessment that takes into account both the increase in soil density and lateral stress produced by ground improvement.

Abstract

This paper describes the experimental methods developed to understand the piping erosion of impervious core and its interaction with downstream filters in embankment dams. Concentrated leaks through a dam core-filter system are simulated experimentally in the hole erosion filter test (HEFT). The mass rate of erosion in the base soil, the eroded soils entrapped in the downstream filters, and the eroded soil particles passed through the filters that settled in the test unit and suspended in the effluent were derived throughout the test. An interpretation method of test results and determination of the rate of erosion of the dam core with downstream filters are proposed for which no methods are reported in the current literature. Results of experimental investigations indicated that the mechanism of erosion in a soil-filter system and the rate of erosion are influenced dominantly by the relative particle size distribution of base soils and filters, applied hydraulic gradient, coefficient of soil erosion, and dry density of base soil. For the soils and filters tested, the average erosion rate of a specimen compacted to 88% of the standard maximum dry density of soil compacted at standard optimum moisture content, Mdd, is about 4.1 times that of specimen with 97% Mdd. The average erosion rate for soil with erosion coefficient $1.54\times {10}^{-5}\mathrm{s}·{\mathrm{m}}^{-1}$ is about 15.4 times of that for soil with erosion coefficient $3.8\times {10}^{-6}\mathrm{s}·{\mathrm{m}}^{-1}$ for the same downstream filter and hydraulic gradient. The average erosion rate for a hydraulic gradient of 27.17 is about 2.2 times that for a hydraulic gradient 18.9 for the same soil and filter tested. Application of the HEFT for assessing the piping risk rating of embankment dams is presented.

Practical Applications

The practical applications of the HEFT presented in this paper extend to the assessment of the piping risk rating of embankment dams. These findings provide a valuable foundation for improving dam safety protocols and the design and maintenance of embankment dams. By gaining a deeper understanding of the factors influencing piping erosion, dam engineers and geotechnical professionals can make informed decisions to enhance the resilience and safety of embankment dam structures, ultimately minimizing the risks associated with this critical geotechnical issue.

The bearing capacity of shallow foundations is often calculated using the classical three-term bearing capacity equation, containing bearing capacity factors and correction factors for embedment depth and footing shape. The available solutions for the determination of these factors were obtained using primarily limit analysis and the method of characteristics, with the assumption that soil is a perfectly plastic material following an associated flow rule. However, in reality, sands are neither perfectly plastic nor do they follow an associated flow rule. The unit bearing capacity of footings in sand depends on (1) the representative friction angle mobilized within the zone encompassing the footing’s collapse mechanism, and (2) the size, shape, and embedment depth of the footing. This paper presents the results of load tests on model strip and square footings of different sizes installed in clean silica sand samples pluviated to produce different densities inside a calibration chamber. The test results for the footings placed on the sand surface show that the bearing capacity factors depend on the relative density of the sand sample and the geometry of the footing. For the embedded footings, the unit bearing capacity of the footing is shown to depend on its embedment depth. Insights into the footing collapse mechanism are provided by plotting the maximum shear strain contours for different stages of footing loading; the contours were obtained using the digital image correlation technique. Finally, we determined the shape and depth factors by analyzing the results of the load tests performed for surface and embedded strip and square footings in dense sand using procedures that account for stress level, loading path, and flow rule nonassociativity.

Three driven precast, four driven cast-in-situ, and four screw injection piles were installed and tested in dense to very dense sand at a site in the Netherlands. Each pile was instrumented with two types of fiber optic sensors and tested under axial compression. Through these tests, a comparison could be made of how different installation methods influence the pile base and shaft response. For example, large residual base stresses were measured in the driven precast piles after installation. Of the three pile types tested, the driven precast piles also reached the highest base stresses, mobilizing their full base resistance at comparatively low displacements. The base response of the driven cast-in-situ piles was also like that of a driven precast pile with residual stresses excluded. In contrast, the screw injection piles mobilized much lower ultimate base resistances and with a much lower stiffness. In terms of shaft resistance, the precast piles showed friction fatigue effects in line with existing models, but this effect was not evident for the driven cast-in-situ or screw injection piles. Finally, shaft and base resistances measured in the dense to very dense sand layers were greater than limiting resistances prescribed in several design standards. By taking this into consideration in design standards, the results would help reduce some of the overconservatism present in design and consequently reduce the financial and environmental cost of pile manufacturing and installation.

Previous plasticity models for predicting anchor kinematics under eccentric loading in clay are mostly rigid plastic. They may produce inaccurate outcomes in some conditions. This paper presents an elastoplastic model to address this issue. Closed-form solutions for anchor velocities are derived considering two types of boundary conditions: the padeye (1) fastened to a mooring chain, with the chain behavior modeled; or (2) hinged to a rigid rod. Compared with previous methods, the proposed method more realistically captures the soil behavior while avoiding complicated iterative procedures for solving anchor velocities, thus enabling anchor trajectories to be accurately and efficiently obtained by simple time integration. A series of 3D large deformation finite element analyses are conducted to study the anchor behaviors under different eccentric loading, and the proposed model is used to reproduce these simulations. Their results are compared for validation, with good agreements achieved. These results indicate that anchors under eccentric loading will ultimately translate along a fixed direction with constant soil reaction forces. Analytical solutions for this ultimate state are given based on the proposed method. Particularly, the ultimate resultant soil reaction force is demonstrated always passing through the anchor center. This is an assumption in a previous literature but is found to be a consequence of the proposed anchor constitutive relations.

Although mechanically stabilized earth (MSE) walls typically require a reinforcement length-to-wall-height ratio of 0.7, the use of short reinforcement may be inevitable in certain situations, such as building MSE walls in mountainous areas, widening existing highways, or restoring eroded roadways. Intuitively, connecting the rear end of reinforcement in shored mechanically stabilized earth (SMSE) walls to the stable face behind can compensate for insufficient reinforcement anchorage length. However, the effectiveness of this approach remains debatable in the existing literature. This study aims to provide more complementary data and insight into the behavior of SMSE walls at a working stress state. A finite difference numerical model was first validated using measurements from centrifuge and full-scale tests. Then, a parametric study was conducted to evaluate the effects of wall geometry, reinforcement parameters, and backfill properties on the lateral facing displacement and maximum reinforcement tensile loads of SMSE walls. The numerical simulations show that, while connected conditions certainly improve the stability of the SMSE wall at the limit state, they may, counterintuitively, worsen the performance of the wall and increase the maximum reinforcement loads at working stress conditions, presenting a design dilemma. The reasons for the results obtained are discussed in detail, and optimized approaches are proposed to improve the performance of SMSE walls.

The drained shear strength of overconsolidated glaciolacustrine clays in the Seattle area can range from the overconsolidated peak to residual strength depending on its geologic history, depth, softening, prior shear displacements, size of excavation, and application, e.g., long-term cut slope versus short-term braced excavation. An inverse analysis of slope failures shows the mobilized shear strength for cut slopes in the overconsolidated and sparsely jointed Qvgl glaciolacustrine clay ranges from the fully softened strength to in between fully softened and residual strength depending on the level of jointing, prior shear displacements, assumed piezometric levels, and Atterberg limits used in the analysis. Conversely, an inverse analysis of a 2011 slope failure in the frequently jointed and random slickensided overconsolidated Qpgl glaciolacustrine clay shows the mobilized shear strength for this deposit near residual to in between residual and fully softened strengths because of repeated glaciation cycles, softening, and local shear displacements due to glacial bulldozing, solifluction, and high lateral stresses.

Building and ground settlement due to construction dewatering is a well-studied topic. However, most previous investigations have not considered the barrier effect of an adjacent underground structure on the drawdowns and resulting settlements. In this study, the barrier effect and its influence during construction dewatering for a metro station foundation pit is investigated. There are five aquifers at the foundation pit site, and a row of buildings supported on pile foundations, which act as an underground barrier to flow, is present on one side of the pit. On the other sides, there are no deep underground structures to impede groundwater flow. Field monitoring of the groundwater level drawdown, diaphragm wall movement, and ground and building settlements on both sides of the pit was carried out during dewatering. The results indicate that on the side with the pile foundations, the groundwater level drawdown, ground settlement, differential settlement, and angular distortion of building incurred by dewatering were relatively greater, but the diaphragm wall movements were relatively smaller. The effect of preexisting barriers should be considered in the assessment of construction dewatering-induced drawdowns, soil settlements, and building movements.

Suction stress is one of the most significant factors affecting the serviceability and stability of soil structures. This study presents a framework for analyzing the stability of geosynthetic reinforced soil (GRS) walls and slopes under unsaturated conditions. An analytical formulation of suction stress-based effective stress was implemented into a limit equilibrium solution, namely, the top-down procedure. The developed framework enables the prediction of the tensile load distribution and connection load between the reinforcement and wall face considering the pullout resistance of reinforcements in GRS walls backfilled with granular and marginal soils under unsaturated conditions. The applicability of the framework was demonstrated by providing an illustrative example followed by three series of parametric studies to understand the effects of pore size distribution, air entry pressure, and infiltration rate on the performance of an unsaturated GRS wall. The results quantify the impact of suction, showing that as it increases, the maximum tensile loads and connection loads decrease while pullout resistance increases. Mostly affected by the suction effect are upper reinforcement layers, where combined effects of reduced tensile load and increased pullout resistance decrease connection load and reinforcement length requirements. The current study strongly discourages counting on the contribution of suction for the design of new GRS walls. The suction value cannot be accurately and reliably determined for the entire lifespan of the GRS wall, and it may decrease or diminish in an uncontrolled and random manner under infiltration. However, by quantifying the effect of suction, the proposed framework in this study provides a valuable tool for analysis purposes, enabling a rigorous interpretation of field-measured reinforcement loads during wall service as well as evaluation of the forensics of failed GRS walls.

Particle crushing takes place in a wide range of engineering applications and affects fluid flow through granular materials. This study examines volumetric changes and the evolution of hydraulic conductivity during particle crushing using a combination of high-stress experiments, data compilation from published studies and particle-scale numerical simulations. An asymptotically correct hyperbolic model fits void ratio versus stress data and provides reliable estimates of the yield stress. Alternatively, hydraulic conductivity follows a power law with void ratio and the trend extends into the grain crushing regime. Hydraulic conductivity diminishes rapidly as sand grains start crushing; the reduction in hydraulic conductivity can reach 1-to-4 orders of magnitude depending on mineral composition, particle shape and initial packing density. Crushing involves the progressive detachment of small fragments; these small particles move into the larger pores, increase the drag resistance, clog pore constrictions, and effectively hinder fluid flow after the yield stress.