Joel Gniel’s research while affiliated with Simplot Australia Pty. Ltd. and other places

Geosynthetic reinforced column supported embankments predominantly utilise two mechanisms to transfer embankment loads towards column heads, soil arching and membrane actions. When undertaking the design of column supported embankments, it is common practice to perform a two-step design, whereby the arching actions are estimated independently of the subsoil deformation and membrane actions. This approach is unable to capture the deformation dependency exhibited by both arching and membrane actions. This paper presents deformation dependent arching and membrane action models and implements them within an interaction diagram. It is shown that an interaction diagram-based design approach is capable of performing an ultimate and serviceability limit state design of a geosynthetic reinforced column supported embankment. In contrast, most existing analytical design methods only consider the ultimate limit state. The proposed method is applied to a design example where the benefits of such a design approach are demonstrated.

The transfer of embankment stresses towards pile heads in piled embankments is attributed to the mechanism known as soil arching. Three-dimensional physical models of piled embankments were built to simulate this mechanism. The progressive settlement of subsoil beneath an embankment was modelled and paused at increments of displacements to allow synchrotron X-ray computed tomography to be performed on the models. Image correlation techniques were then applied to the reconstructed volumes to obtain evolving three-dimensional displacement and strain fields. The strain fields show localised (shear bands) and diffuse failure modes occurring above pile heads within the embankment fill. These failure surfaces are seen to progressively develop as the subsoil undergoes settlement. The displacement fields also show the formation of a plane of equal settlement developing at a height above the pile heads, known as the critical height. The critical height is dependent on the height at which the failure surfaces propagate into the embankment fill, and a method is proposed to calculate the maximum height of failure surfaces based on the observed kinematics. The full-field kinematics provide fundamental insight into the soil arching mechanism that develops within piled embankments.

For geosynthetic reinforced column supported embankments (GRCSE) supporting a high embankment, lateral forces associated with lateral sliding and embankment stability often govern the acceptability of a given design under serviceability conditions. Frequently, the complex soil–structure–geosynthetic interaction, the size, and the three-dimensional nature of a GRCSE necessitate the use of numerical analysis to assess embankment performance relative to serviceability criteria. However, traditional finite element method techniques used to model serviceability behaviour are limited in their ability to model the geotechnical mechanisms associated with column installation, equilibration, and group installation effects. These installation effects are examined herein based on a GRCSE field case study located in Melbourne, Australia, that has been extensively instrumented. The role that these installation effects have on the performance of the GRCSE is highlighted and the behaviour of the columns supporting the embankment is emphasized. It is shown that cracking of the unreinforced columns supporting the embankment is likely inevitable and that the reduction of lateral resistance provided by the columns should be accounted for in design. The suitability of various numerical approaches currently used in design to model the columns supporting the GRCSE, and the embankment itself, are discussed and recommendations are made.

Post-construction data from an instrumented geosynthetic reinforced column supported embankment (GRCSE) on drilled displacement columns in Melbourne, Australia, show the time-dependent development of arching over the 2 year monitoring period and a strong relationship between the development of arching stresses and subsoil settlement. A ground reaction curve is adopted to describe the development of arching stresses and good agreement is found for the period observed thus far. Predictions of arching stresses and load-transfer platform behaviour are presented for the remaining design life. Four phases of arching stress development (initial, maximum, load-recovery, and creep strain phases) are shown to describe the time-dependent, and subsoil-dependent, development of arching stresses that can be expected to occur in many field embankments. Of the four phases, the load-recovery phase is the most important with respect to load-transfer platform design, as it predicts the breakdown of arching stresses in the long term due to increasing subsoil settlement. This has important implications in assessing the appropriate design stress for the geosynthetic reinforcement layers, but also the deformation of the load-transfer platform in the long term.

In recent years, geosynthetic reinforced column supported embankments (GRCSEs) have become an increasingly popular design solution for road and rail infrastructure constructed over soft soil sites. However, the serviceability behaviour and deformation that often govern the suitability of their design is not well understood. This is due, in part, to the difficulties in describing the arching stress development in the load transfer platform (LTP). This paper highlights the need for coupled arching stress-deformation models to describe accurately serviceability behaviour. This approach contrasts the widely adopted two-step design approach, which uses limit-equilibrium models that de-couple the arching stress-deformation relationship to describe ultimate limit state behaviour. Using an analytical example, an arching stress/deformation model and an empirical relationship (developed by others) relating base LTP settlement to surface settlement, the relationship between serviceability behaviour and soft soil parameters is highlighted and the conditions leading to progressive collapse in GRCSEs are described. The approach presented provides a means to predict serviceability behaviour, and at the same time, raises questions about the long-term performance and the manner in which acceptable performance has been achieved in the short-term in several field case studies. In particular, those constructed at, or near, a minimum embankment height.

As part of the Regional Rail Link (RRL) project in Melbourne, Australia, a number of geosynthetic reinforced (GR) embankments with ground improvement were constructed where the rail alignment passes over the Coode Island Silt (CIS), a well-known soft soil encountered around inner Melbourne. To better understand the behaviour and performance of the load transfer platform (LTP) at the base of these embankments, a field case study has been undertaken which has seen an extensive array of instrumentation installed within the North Dynon embankment. This paper presents and describes a significant amount of field and laboratory data gathered as part of the geotechnical site characterisation of the instrumented areas. Based on this data a description of the compressibility and permeability of the CIS is presented and insight into the structured nature of the CIS is described. In addition, it is shown that a far better characterisation of soft soil behaviour can be gained through the use of more sophisticated oedometer testing techniques.

In this study, an interconnected design-chart is developed to design prefabricated vertical drain (PVD) improved ground taking into account the effect of the following factors on the radial consolidation rate where the case of uniform subsoil is assumed: (i) soil permeability and compressibility changes during the consolidation process, (ii) transition smear zone with hydraulic conductivity varying linearly, (iii) time dependent loading. The Finite Element Method (FEM) was used to solve the general governing radial consolidation equation where the above mentioned factors are included. The obtained FEM solution was validated using laboratory test results and analytical solutions for certain cases in the literature. Then FEM solution was used to develop PVD design charts that incorporate the effect of the above mentioned factors.

A recent rail infrastructure project in Melbourne, Australia has seen the construction of a number of geosynthetic reinforced (GR) piled embankments to overcome the challenges associated with the local soft soil present throughout much of inner Melbourne, Coode Island Silt (CIS). To better understand the behaviour and performance of the load transfer platform (LTP), a field case study has been undertaken involving the instrumentation of one of these embankments. This paper presents data gathered during the construction and post-construction phases (approx. 9 months) as well as discussion on the LTP behaviour observed thus far. The development of arching, as seen from these preliminary results, is currently increasing with time; it is suggested that this increase is related to a coupled response between the arching mechanism and the GR/sub-soil deformation. The effects of piling works in close proximity to the LTP are also discussed and further work as part of this on-going research is described.

Geogrid encasement has recently been investigated to provide an alternative and perhaps stiffer option to the now established method of geotextile encased columns (GECs). To construct geogrid encasement, the geogrid is typically rolled into a sleeve and welded using a specialized welding frame. However, the process is unlikely to be economical for site construction and therefore an alternative method of encasement construction was investigated in this paper. The technique comprises overlapping the geogrid encasement by a nominal amount and relying on interlock between the stone aggregate and section of overlap to provide a level of fixity similar to welding. A series of small-scale tests were initially used to investigate the technique, followed by medium-scale compression tests using different geogrids and typical stone column aggregates. The results of testing indicate that the “method of overlap” provides a simple and effective method of encasement construction, providing a level of fixity similar to welding. A full circumference of overlap should generally be adopted to achieve adequate fixity. Biaxial geogrids are best suited to the technique, with increased encasement stiffness resulting in increased column capacity and column stiffness. Higher strength geogrids are also more robust, providing a greater resistance to cutting from pieces of angular crushed rock. Site trials are recommended for final confirmation of the technique.