Keywords

1 Introduction

Ground models are required over the full course of planning, design and construction of linear infrastructure projects. The typical progression of a ground engineering project would see the level of detail and certainty of the ground model increase over the course of the project as the ground model develops in parallel with the project and design development. The uncertainty within the ground model would then generally reduce in parallel with the progression of the project. A challenge facing any geotechnical engineering project is in identifying how much uncertainty can be tolerated at any stage of the project. An assessment of the uncertainty in the ground model and its associated risks informs the investigation that needs to be undertaken over the course of the project and what risks must be accommodated in design.

Amongst other things, the usefulness of ground information used as input for developing a ground model depends on:

  • the reliability that can be placed on the ground information;

  • the quantity of information and its location or relevance to the project for which it is being assessed;

  • the geological complexity and variability of the ground at the location of interest, and;

  • the susceptibility of the project to ground conditions.

Assessing how much investigation should be undertaken at various stages throughout the project, requires an assessment of how much ground model uncertainty can be tolerated. A method is set out below which seeks to qualify the uncertainty associated with a ground model for linear infrastructure projects which can be applied throughout its various stages. The method proposed follows the process as set out in Fig. 1. Two key streams are considered: the available ground information (left column on Fig. 1) and the ground complexity (right column on Fig. 1). An assessment of each is undertaken and then qualitatively combined in order to assess ground model reliability.

Fig. 1
figure 1

Flow chart indicating steps in assessing ground model uncertainty. Numbered steps may be cross referenced to headings in this paper

Full size image

This paper describes a qualitative process to assess ground model uncertainty which is similar to and draws inspiration from qualitative systems used to assess risk, for example the RTA system of slope risk assessment, Stewart et al. (2002). Each step of the qualitative process is described in this paper. The numbered headings may be cross referenced to the numbers shown in the flow chart in Fig. 1.

2 Assess the Reliability of Available Geotechnical Information

The ground model at the concept or feasibility stage of a project is likely to be based on a desk study and informed by existing information that is unlikely to have been acquired specifically for the project. It may have been acquired to inform geotechnical aspects of other projects or for entirely unrelated, non-geotechnical purposes and therefore have varying relevance. For urban infrastructure projects for which the method set out in this paper was developed, the available information was predominantly historical borehole information and a method of assessing the quality of borehole information is set out here. Notwithstanding this, the methods set out here for boreholes could be adapted for map information, CPT, geophysics or any other relevant information.

In order to assess the reliability of boreholes, a semi-qualitative assessment can be made on a borehole by borehole basis. The method suggested here uses a similar approach to that used in rock mass classification systems such as RMR (Bieniawski 1989). Various attributes of the borehole are assessed and score is allocated for each attribute as indicated in Table 1. The scores for each attribute are then summed with the score allocated to a borehole reliability category, 1–5. Although development of the borehole reliability score is semi-quantitative, it is important to note that the overall system presented here is qualitative and that the borehole reliability score is used to provide a qualitative input (borehole ranking) to the system. The attributes considered for boreholes are set out below.

Table 1 Categories used to assess borehole reliability score
Full size table

Drilling Method—Boreholes with core drilling including an assessment of rock defects are more reliable than boreholes using washbore or hammer techniques.

Survey—Boreholes with recent ground survey are considered more accurate than those without survey and allow more reliable positioning in ground models.

Sampling Frequency—The greater the sampling frequency in the borehole, the more reliable the soil or rock description is considered to be.

Age—There is greater uncertainty around the provenance of older borehole logs.

Depth—Shallow boreholes are generally considered less relevant than deep boreholes, particularly for tunnel projects.

Installation—Boreholes with groundwater wells installed and groundwater measurements are considered more useful than boreholes without well installations.

In situ Testing—Boreholes with in situ testing including SPTs, packer testing or pressuremeter testing provide information on engineering properties of the ground.

Note that the above criteria have been developed with a focus on urban tunneling projects but could be readily adapted to other forms of linear infrastructure projects or other types of projects.

The process of selecting suitable numbers for each category was one of trial and error. A borehole reliability assessment was undertaken for approximately 1000 boreholes and the scores assigned to each category varied until the output was considered reasonable. Judgement was exercised in undertaking this trial and error process. If required, it is a relatively simple process to vary the categories and numbers to include project or location specific attributes and weightings.

The reliability scores are then assigned to a category in accordance with Table 2 which can then be used as input to the qualitative assessment. Similar to the parameters used to develop the reliability score, the category ranges have been developed through a trial and error process and can be tailored to the specific project and location. The implications of this borehole reliability ranking to future investigation and ground model development are set out in Table 2.

Table 2 Borehole reliability ranking
Full size table

Boreholes ranked 1–3 would generally be used to inform ground model development, supplemented by boreholes with a lower ranking. Figure 2 provides an indication of the borehole reliability spatially along part a project assessed using the methods set out above.

Fig. 2
figure 2

Example of borehole reliability plotted on portion of linear infrastructure project

Full size image

3 Assess Quantity and Quality of Data Spatially Relative to the Project Alignment

To assess information quality and quantity, the borehole reliability ranking can be used in conjunction with an assessment of the density of boreholes relative to the proposed project alignment. Table 3 provides an example of how this was undertaken for an urban tunneling project in Melbourne, Australia where the assessment was undertaken over 100 m intervals along the proposed tunnel alignment.

Table 3 Borehole information quality and quantity
Full size table

4 Assess Geological Complexity

Geological complexity refers to ground characteristics with complexity related to aspects such as variability of lithology, complexity of structure, stress field and tectonic history. It might also be described as the degree of ground homogeneity at the scale of the project.

Table 4 presents a guideline for the assessment of geotechnical and geological complexity which was developed for an urban tunneling project in Melbourne.

Table 4 Assessment of geological complexity at scale of tunneling project
Full size table

For example, the geological complexity of normally to slightly overconsolidated Holocene sediment may be ‘simple’ because its depositional history and stress history are known and its engineering properties are relatively predictable. Conversely, a metamorphic rock mass with multiple phases of deformation and weathering would be ‘complex’. Note that geological complexity should be assessed at a scale relative to the projects zone of influence.

5 Assess Geotechnical Susceptibility

Geotechnical susceptibility refers to the susceptibility of the project or project element to ground uncertainty.

Table 5 presents a guideline for the assessment of geotechnical susceptibility which was used to qualitatively assess an urban tunnelling project in Melbourne.

Table 5 Assessment geotechnical susceptibility
Full size table

Table 6 presents an example of the geotechnical susceptibility that might be estimated for various elements of an urban metro project.

Table 6 Estimated geotechnical susceptibility of various elements
Full size table

6 Combine Assessments of Geological Complexity and Geotechnical Susceptibility to Assess Ground Complexity

The assessments of geological complexity are combined to arrive at an overall estimate of ground complexity. Table 7 provides a matrix which combines geological complexity and ground susceptibility. This matrix is biased towards geotechnical susceptibility. For example complex geological complexity and intermediate geotechnical susceptibility combines to intermediate.

Table 7 Tool to assist in estimating ground complexity based on geological complexity and geotechnical susceptibility
Full size table

7 Combine Assessments of Ground Information and Ground Complexity to Qualitatively Assess Ground Model Uncertainty

The assessments of ground information (step 3) and ground complexity (step 6) are combined using the matrix in Table 8, to arrive at an overall ground model reliability ranking. An example of how the ground model reliability may be communicated on a ground model, in this case a simple cross section, is presented in Fig. 3. The ground model uncertainty was superimposed over the ground model at project feasibility stage using desktop information.

Table 8 Ground model reliability rating
Full size table
Fig. 3
figure 3

Example of cross section presenting ground model reliability for a tunneling project. Ground complexity for each geotechnical unit and ground information are combined to indicate Ground Reliability (lower section)

Full size image

8 Identify Implications of Ground Model Uncertainty

An indication of the implications of the estimated ground model reliability score is provided in Table 9.

Table 9 Implication of ground model reliability score
Full size table

This approach informs an assessment of where further investigation might be expected to provide the most value and can be refined multiple times as the project progresses and additional information becomes available.

9 Conclusions

A method is set out here by which the reliability of a ground model may be assessed in a semi-quantitative manner. The methods described have been applied with success on three major tunneling projects in Melbourne, Australia. However, the methods described here are adaptable. The criteria and weightings assigned to criteria for borehole reliability assessment can be modified through a trial and error process and tailored to a specific project and geological setting.

Whilst the question of how much investigation is enough investigation cannot be answered directly using the technique described here, this method may inform identification of those areas where further investigation will add the most value.