2020 Young Geotechnical Professionals’ Night

The current accepted practice for road earthworks compaction quality assurance in NSW is a statistical conformance approach using standard density and moisture ratio. This method of compaction quality assurance has been accepted practice for over 25 years. The use of machine-based intelligent compaction systems in NSW has been minimal to date, mainly due to historic poor correlation with density and moisture ratio. As part of the Wells Crossing to Glenugie 2a Pacific Highway upgrade project, Lendlease Engineering have undertaken a project-wide trial of intelligent compaction. Conducted in tandem with the contract earthworks specification, this trial aims to understand both the earthworks efficiency and potential quality assurance benefits with respect to current practice.

The trial has focused on upper zone of formation layers including the select material zone (UZF and SMZ). The intelligent compaction testing includes portable lightweight falling deflectometer testing (LWFD) and flat drum roller accelerometer-based intelligent compaction (CMV). The testing is conducted and analysed in combination with survey position and level technology installed in the placement and compaction fleet.

Results to date show an improved earthworks placement and compaction methodology with greater lot performance traceability. The trial has shown a variable correlation with density ratio and moisture ratio with respect to material type. Stronger correlations are developed between the machine-based CMV, LWFD and Benkelman Beam testing.

Conclusions test the relevance of density and moisture ratio as well as testing frequency within current specification. The study considers replacing or supplementing testing with intelligent compaction for future earthworks specification.

The Tauranga Crossing Stage 2 Development comprises an AUD $150 million, 12,000m2 extension to an existing multi-storey retail centre. Ground conditions across the site are complex and include sensitive and reworked volcanic soils, interbedded alluvial deposits, with frequent organic swamp and peat layers also encountered throughout. A central geological feature of the site is a wide estuarine paleochannel within the building footprint, found to extend up to 12 metres below ground level, and containing very soft, heterogeneous in-fill deposits with undrained shear strengths typically less than 30 kPa.

Initial development and preloading of the site was commenced in 2006, inducing over 1 metre of settlement as of 2015, but was not well aligned with the final layout of the development and, arguably, introduced an additional level of complexity with respect to site characterisation and design. Coupled with New Zealand’s high seismicity environment, the site conditions presented numerous geotechnical hazards, including a high potential for seismic liquefaction and cyclic softening, primary consolidation, soil creep and differential settlements, respectively.

Ground improvement was identified as the optimal means of progressing site development and a performance-based specification was released for procurement. Golder Associates were awarded the design & build contract, with the tender design comprising Rammed Aggregate Pier® elements (RAPs); a technique patented by The GeoPier Foundation Company. Owing to the low lateral confinement expected in-situ, the RAP design specification included a weak grout during installation to stiffen elements beneath heavily loaded shallow footings, and to prevent a bulging type failure at low confining stresses. At the time of construction, this was the first known application of grouted RAPs in New Zealand, remaining cost-competitive compared to other ground improvement techniques considered.

This paper outlines the design requirements and rationale, as well as the design optimisation afforded by undertaking trial RAPs, additional geotechnical and geophysical investigation, laboratory testing, load testing and coring of completed RAPs as construction progressed. This allowed the RAPs to be curtailed in some areas and by up to 5 – 6 metres compared to the tender design assumptions. Construction verification culminated in a full-scale load test, and which re-affirmed the efficacy of the scheme in satisfying the performance requirements; the results of this test are discussed and compared retrospectively with original design calculations and load tests upon individual RAP elements.

The use of deep segmental casing in lieu of sacrificial liners clearly offers a material saving for the steel liners, and is beneficial in reducing programme and obstruction risk. However, there are technical challenges related to its use and it is important that the Piling Contractor is suitably experienced in their use. These include the need for large heavy-duty piling rigs with high rotary torque and high line pull capacity as well as the careful measurement of concreting levels during concrete placement and casing removal. It was identified that the concrete loss at certain depths is more and less at other depths on the same site. This raised the question about the relationship between concrete loss and layers of soil. In order to narrow down to reasonable findings, an assumption was made, that pile diameter will not affect the amount of concrete loss through a layer. Furthermore, in secant pile wall construction it was assumed that hard piles will have a relatively smaller amount of concrete wastage than soft piles. The presenter will deliver an evidence-based approach that clearly identifies and explores the rationale for this. We will also discuss how various soil layers affect the concrete quantities that need to be placed to assure a monolithic and conforming pile shaft.

Assessing the load is one of the first steps in tunnel lining design. While there are often a combination of loads acting on the lining such as groundwater load, live load, rock load and M&E load etc., this paper focuses on the ground load. Three categories of methods for estimating rock loads are discussed: graphical, empirical and numerical modelling.

In the graphical method, the proposed tunnel shape is placed at random locations within a network of geological defects representing the expected ground conditions. Potentially unstable blocks are identified around the tunnel crown perimeter. The equivalent rock pressure is calculated from the weight and size the block intersecting the tunnel perimeter. The difficulty lies in producing a model sufficiently representative of the ground, in terms of defect spacing, persistence and orientation from available site investigation data and designers’ experience in the regional and local geology. The Discrete Fracture Network (DFN) method is covered under this category. It looks at probability of occurrence for blocks based on a 3D geometric model of defects distributed in a rock mass.

The paper presents some of the most cited empirical methods. The advantages and limitations of several methods are outlined, along with interpretation of key parameters in the context of Sydney sandstone.
Numerical analyses are employed to simulate the load transfer onto tunnel support. The analyses can be staged to include in-situ stress initialisation, sequential excavation, installation of primary support, and permanent lining. The stresses on the lining can be extracted as ground load from the numerical model. The rock mass can be modelled either as a homogeneous elasto-plastic medium or as intact rock with discrete defects. The former gives uniform results but often overestimates the load. The latter can lead to spikes and outliers which require smoothing in post-processing. The magnitude of ground load redistribution and loading sharing between primary rock bolts and the lining depends on ground conditions, in-situ stresses and stiffness of the lining and interface element. Parametric studies are often required to bound the resulting ground load. The interaction of adjacent excavations, such as at tunnel intersections or crossovers, typically requires three-dimensional analyses. The comparison of rock loads derived using graphical, empirical and numerical methods are presented to demonstrate the application of the various methods in a practical setting.