1   Introduction

Soil conditioning is an important part of the tunnelling process with Earth Pressure Balance (EPB) TBMs. It is used to modify the properties of the excavated soil to improve control of the excavation process and the TBM performance. Conditioning agents such as water, foam and polymers are injected to the soil during excavation to ideally form a soft, plastic paste with low permeability that supports the tunnel face and flows through the TBM in a controlled manner. The properties of the conditioned soil that best suit the TBM operation can be quite different to those required for transport and disposal or re‐use of the spoil generated by a tunnelling project. Environmental regulations also play a role in the use of soil conditioning for EPB tunnelling projects. The design and application of soil conditioning treatments should consider these different stages and requirements to manage the excavated spoil as effectively as possible.

The conditioning treatments required to improve the soil properties for EPB tunnelling can vary significantly for different ground conditions. There are also a wide range of products available from different suppliers. Laboratory testing of conditioned soil samples prior to tunnelling and of TBM spoil samples during construction provides information for selection of conditioning treatments and monitoring of the tunnelling process.

This paper first describes different laboratory test methods for soil conditioning and their role in an EPB tunnelling project. The testing performed on the Channel Tunnel Rail Link and Crossrail projects in London, UK, and the Port of Miami Tunnel in Florida, USA, is then described to illustrate applications for a range of ground conditions as part of the construction planning and excavated soil management for these projects.

2   Laboratory testing for EPB soil conditioning

2.1 Purpose of testing

Preliminary assessment of the feasibility of EPB tunnelling and soil conditioning treatments for different ground conditions can be made from geotechnical information routinely gathered during project planning and design stages. Particle size distributions, Atterberg limits, permeability and in‐ situ groundwater pressures can be used with guidance such as that presented by Thewes (2007), DAUB (2010) and EFNARC (2005). Information from other projects in similar ground conditions can also provide experience and reference data.

Laboratory testing of conditioned soils prior to tunnelling can provide more detailed information for project‐specific ground conditions to assist with construction planning. Testing of TBM spoil samples during tunnelling can also be used to assist with monitoring the construction and spoil transport / disposal processes.

Purposes of these tests include:

• technical evaluation of different conditioning agent products
• characterisation of sample properties to determine effective conditioning treatments and expected TBM spoil properties for transportation and disposal / re‐use
• assessing the feasibility of effective conditioning and EPB tunnelling in difficult ground
• assessment of effective conditioning parameters and agent consumption rates
• environmental testing to inform impact assessments and approval procedures for use of conditioning agents in tunnelling and for disposal / re‐use of conditioned spoil
• monitoring TBM spoil properties to optimise conditioning and TBM operations, and for transportation or disposal requirements.

2.2 Soil conditioning laboratory test methods

There are various factors to consider in laboratory testing of conditioned soils, including:

• the soils to be encountered and proportions of different soils within the tunnel face;
• the effect of the excavation process on the composition and properties of the soil(s)
• the various types of conditioning agents, products and treatment parameters;
• the properties of the conditioned soils relevant to the excavation and spoil transport /
  • disposal processes;
• environmental testing and approval requirements.

There are no formal standard test methods for conditioned soils. However, various index tests are commonly used in research and practice for measuring basic properties of conditioning agents and conditioned soils, including:

• Foam expansion ratio and stability (liquid drainage time)
• Slump tests to measure the ‘plasticity’ or ‘workability’ of conditioned granular soils
• Shear vane tests to measure undrained shear strength of conditioned soils
• Mixing tests to observe consistency and stability of conditioned soils, possibly with measurement of power consumption or torque during mixing
• Adhesion test to measure conditioned soil‐steel interface friction / adhesion of clays
• Mortar flow table test to measure ‘plasticising’ effect of foam on a soil

These tests are relatively simple and fast to perform (usually at atmospheric pressure), and are useful for comparing the relative performance of a range of conditioning treatments for different soils and the effects on sample properties. They are well suited for laboratory test programmes prior to tunnelling to inform the selection and application of conditioning treatments for project‐specific soils, and to investigate variables that influence the conditioned soil properties.

Index tests are also suitable for routine on‐site testing of TBM spoil properties as part of construction monitoring and management of the spoil transport and disposal processes. It is important to note that the properties of conditioned soils can vary over time and throughout the construction process, so on‐site sampling and testing should be undertaken at appropriate points. Test samples for spoil properties relevant to TBM operations can be taken close to the screw conveyor outlet. Samples for spoil transport or disposal tests can be taken from a spoil conveyor belt or stockpile, depending on the project‐specific processes.

Specific tests or criteria may be required depending on the local regulations applicable to the spoil transport method or the spoil disposal / re‐use location or application for a project.

More advanced laboratory tests are also used to measure fundamental properties of conditioned soils and their influence on the EPB excavation process under more realistic pressurised test conditions. These methods for element tests and physical model tests were often developed for research applications but are also used for project‐specific testing, and include:

• Permeability tests on conditioned granular soils
• Shear box tests on conditioned soils and soil‐steel interface shearing
• Compressibility tests on conditioned soils under varying pressures
• Pressurised mixing tests with measurement of torque or power consumption
• Pressurised shear vane and adhesion tests for conditioned clay soils
• Laboratory scale model EPB screw conveyor tests

For project‐specific testing, these types of advanced laboratory tests are less suited for assessment of a wide range of conditioning treatments or for construction monitoring. However, they can be used to better assess the properties and performance of soils conditioned with treatments selected based on an initial index test programme. They can also be used to investigate the feasibility of EPB tunnelling in difficult ground where more realistic test conditions are appropriate.

To plan and interpret laboratory tests performed to assess the properties of conditioned soils, it is important to have information on the soil properties such as the composition, Atterberg limits, in‐ situ water content and groundwater pressures, permeability and particle size distributions. Consideration should be given to the effects of the excavation process on the soil properties such as the grading of a soft rock or cemented material after excavation, or the influence of the EPB face pressures or external dewatering on the water content of the excavated soil. These factors should be considered to prepare conditioned soil samples that are representative of the excavated materials produced by a TBM.

Conditioning agent consumption rates estimated from laboratory tests can be used in environmental testing and assessments for the use of specific conditioning agents according to local regulations. These requirements vary but can include chemical testing of the concentrated conditioning agents for potential contaminants, or testing based on the expected consumption rates using methods appropriate for the spoil management plan.

Laboratory soil conditioning tests allow a more informed approach to determining effective treatments than trial and error during tunnelling or reliance on previous experience. Conditioning parameters are then adjusted and optimised during tunnelling based on the TBM performance and the ground conditions encountered. Testing of spoil samples during tunnelling can assist with optimising the conditioning treatments and with management of the spoil transport and disposal process.

The soil conditioning and TBM spoil testing performed for some EPB tunnelling projects are described in the following sections.

3   Channel Tunnel Rail Link (CTRL), London, UK

Contract 220 of the CTRL project involved construction of 7.5km of twin‐bore, 8.15m diameter tunnels using two EPB machines from Stratford to St Pancras between 2002 and 2004. The tunnel depths ranged from 5 to 40m below ground level and passed through the soils typically encountered in London including the stiff high plasticity London Clay, the variable stiff intermediate to high plasticity silty clay and silty, sandy gravel beds of the Lambeth Group, the Thanet Sand (silty fine sand), and the Chalk. Mixed face conditions were encountered along significant tunnel lengths.

A laboratory testing programme was undertaken by the Contractors in collaboration with a research project at Cambridge and Oxford Universities to evaluate different soil conditioning agents and treatments for the soils encountered on the project. The testing included:

• Index tests to compare properties of five foam agents from three different suppliers
• Shear vane and fall‐cone tests to measure the undrained shear strength of London Clay and
  • Lambeth Group samples conditioned with foam and polymer treatments
• Slump tests to assess foam and polymer conditioning treatments for Thanet Sand
• Laboratory model EPB screw conveyor tests to measure the effects of conditioning treatments on the operation and control of the excavation process.

The index tests showed that foam conditioning treatments using typically recommended parameters did not perform well with the high plasticity London Clay, as the liquid phase of the foam was absorbed by the soil leading to rapid breakdown of the foam. The best results were achieved using combined foam and polymer solution conditioning treatments, which improved the performance and stability of the foam and reduced the shear strength of the conditioned clay to form a soft, plastic paste suitable for EPB tunnelling. The total conditioner liquid injection ratio (LIR), determined from the foam injection ratio (FIR) and expansion ratio (FER) and the polymer injection ratio (PIR), was a key parameter influencing the strength of the conditioned London Clay, as illustrated in Figure 1. From the results of these index tests, conditioning treatments and parameters required to achieve particular strengths of the conditioned London Clay were determined (values in the range of 10 to
25 kPa have been suggested as suitable). Further details are presented in Merritt et al (2003).

Slump tests on Thanet Sand samples showed that effective conditioning could be achieved using foam or water‐based polymer solutions, with the sample properties dependent on the conditioner injection ratios, polymer concentration and the water content of the soil (Pena, 2003).

The model EPB screw conveyor tests showed that controlled operation could be achieved for London Clay and Lambeth Group samples conditioned with foam and polymer treatments. The material flow rates along the conveyor were controlled and the pressure gradient was stable. The torque or power required to operate the screw conveyor was directly related to the undrained shear strength of the samples and the conditioning treatments. Further details of these tests are given in Merritt and Mair (2006).

The soil conditioning agents and treatments used during tunnelling were selected based on results from the laboratory tests. The conditioning plant installed on the TBMs was also modified to allow independent injection of foam and polymers or other conditioning fluids. Combined foam and polymer solution conditioning treatments were used in all soils, although with different FIR, PIR and total LIR values. More foam and less liquid were used in the Thanet Sand, whereas less foam but more liquid was added to the London Clay through a higher PIR and lower FER (i.e. a ‘wetter’ foam containing more liquid). The conditioning treatments used during tunnelling are summarised in Figure 2 as average and standard deviations of injection parameters used for different soil types. Further details of the effects of soil conditioning on the performance of the TBMs based on field monitoring during tunnelling are given in Borghi and Mair (2008).

The conditioning treatments used during tunnelling and their effects on the excavated soil properties were generally consistent with those assessed from the laboratory tests and allowed controlled EPB tunnelling in the varying ground conditions. Conditioning of the high plasticity clays was achieved by injecting large volumes of liquid to form a soft paste; however, a large power input was required to excavate at slower advance rates in these soils. The spoil water content was also increased significantly and additional treatment was required for final spoil disposal, increasing costs overall construction costs.

4   Crossrail, London, UK

The Crossrail project includes construction of 42km of 7.1m diameter running tunnels for a new railway running east‐west beneath central London. Two of the three tunnelling contracts (covering Drives X, Y, Z and G) are using 6 No. EPB machines to construct 36km of tunnels through mainly London Clay, with some sections in the Lambeth Group and Thanet Sands. Further background on the Crossrail project and management of the excavated spoil transport and disposal / re‐use are given in Davis and Russell (2014).

4.1 Soil conditioning tests for London Clay

Recognising the difficulties with effectively conditioning the stiff, high plasticity London Clay for efficient TBM operations and spoil management, an independent soil conditioning test programme was commissioned by the Contractors prior to tunnelling.

Index tests including mixing tests with measurement of power consumption and shear vane tests were performed on samples of London Clay cuttings obtained from shaft excavations, mixed with a wide range of conditioning treatments. The test programme included 19 different foam and polymer conditioning agents from 6 different suppliers, who had proposed the products based on their own testing and/or experience. Representatives from the suppliers were involved with the testing of their products. The objectives of the test programme were:

• Technical evaluation of the different conditioning agents and treatments
• Condition the soil to reduce the shear strength, stickiness and mixing energy while minimising the amount of conditioning liquids used and the spoil water content
• Estimation of product consumption rates to optimise the conditioned soil properties

Results from the laboratory index tests performed prior to tunnelling are shown in Figure 3 as the water content and vane shear strength of the conditioned London Clay samples. Also shown are data from TBM spoil tests during tunnelling discussed in Section 4.2.

The laboratory tests with a wide range of conditioning treatments produced samples with shear strengths ranging from approximately 5 to 40 kPa with water contents between 27 and 47%. There were two main approaches to conditioning the London Clay recommended by the different suppliers:

Figure 3  – Shear strength of conditioned soils from laboratory and TBM spoil testing.

1. Using foam agents containing dispersant additives to reduce the stickiness of the clay, with FIRs typically between 80 to 100% and an additional 5 to 10% water to improve the foam performance. These treatments added 10 to 15% liquid, producing samples with water contents of about 30 to 40% and strengths of 15 to 30 kPa.

2. Using foams at FIR 50 to 60% in combination with a water‐based polymer solution at 20 to 30% injection ratio to improve the foam performance and provide lubrication to reduce the clay stickiness. These treatments added 30 to 35% liquid, producing samples with water contents of about 40 to 45% and strengths of 5 to 12 kPa.

FERs of about 8 to 10 were typically recommended by the suppliers for the testing. By adjusting the FIR, FER and water / polymer solution volume, similar total liquid injection ratios can be achieved using different conditioning treatments.

Product consumption rates were estimated for the recommended conditioning treatments which allowed commercial comparisons to be made for their use during tunnelling. Foam agent consumption rates typically varied between 1.0 and 3.0 l/m3 (litres of foam agent per unit volume of excavated soil), and polymer consumption rates between 0.15 and 0.3 l/m3.

Based on the preferred conditioning treatments and product consumption rates, environmental risk assessments for soil conditioning during tunnelling in the protected deep groundwater aquifer in London were undertaken to gain approval for the use of these chemicals. Environmental laboratory testing of the conditioning agents was also undertaken according to the requirements for disposal of the tunnel spoil at Wallasea Island.

4.2 Soil conditioning during tunnelling and TBM spoil testing

The bulk of the excavated material produced by the Crossrail tunnelling works is transported by rail and ship to the final disposal site at Wallasea Island. As discussed in Davis and Russell (2014), this transport process imposes some limits on the TBM spoil properties and regular testing is performed on samples taken from spoil stockpiles to measure the water content, Transportable Moisture Limit (TML) and undrained (vane) shear strength of the spoil prior to acceptance for shipping. These testsalso allow comparison of the TBM spoil properties and conditioning treatments with the laboratory tests performed prior to tunnelling.

The soil conditioning used during tunnelling different sections of Drives X, Y and Z (currently under construction) are summarised in Figure 4. For a 320m length of Drive X in London Clay, an average FIR of about 30% was used with a liquid injection ratio of 10%. For a 600m length of Drive Y in London Clay, significantly higher foam and liquid injection ratios were used. The different drives used foam agents including clay dispersant additives from different suppliers, with low FERs of about 1.5 to 3.0 indicating that the liquid component principally conditioned the high plasticity clay. For these sections of Drives X and Y in London Clay the foam agent consumption rates were approximately 2.2 to 4.2l/m3. The different conditioning used for London Clay reflects different approaches to the TBM operations by the Contractors.

Different conditioning treatments were used during tunnelling 2300m through the Lambeth Group and 1825m through Thanet Sand on Drives Y and Z. The Lambeth Group encountered is highly variable with an intermediate to high plasticity depending on the composition. More foam and less liquid was used to condition the Thanet Sand, which was often encountered in a mixed face with the overlying Lambeth Group; these are more granular soils but can have an intermediate plasticity depending on the amount of fines present. Foam agent consumption rates for these sections of the tunnels were approximately 2.5 l/m3 for the Lambeth Group and 1.5 l/m3 for Thanet Sand.

Properties of the TBM spoil are summarised in Figure 3 for high plasticity (liquid limit >50%) and intermediate plasticity (liquid limit 20 to 50%) samples from Drives X, Y and Z. The Drive X London Clay spoil samples had strengths of about 20 to 40 kPa, similar to the laboratory samples conditioned with similar amounts of liquid to water contents of 30 to 40%.

The properties of some of the high plasticity spoil samples from Drives Y and Z are similar to those from Drive X and the laboratory tests; however, the results are more variable with many spoil samples having higher water contents than the laboratory samples. These samples were taken from the mixed spoil from two TBMs operating in different London Clay and Lambeth Group soils, and it is difficult to accurately determine their composition.

The range of spoil properties probably reflects the higher liquid injection ratios used during tunnelling (compared to Drive X) and the variability of the ground, in particular the presence of water bearing sand channels in the Lambeth Group. The intermediate plasticity samples from Drives Y and Z had significantly lower strengths and water contents, reflecting their more granular composition from the lower Lambeth Group and Thanet Sand deposits.

The results of the laboratory index tests prior to tunnelling are reasonably consistent with the TBM spoil sample properties, particularly for Drive X which are the most directly comparable. While the conditioning parameters used for tunnelling in London Clay differed between the TBMs and the laboratory tests, the volume of liquid injected covers a similar range and appears to strongly influence the strength of the conditioned soil.

4.3 Spoil properties for tunnelling and disposal

As noted in Davis and Russell (2014), a minimum undrained shear strength of 20 kPa was required for safe shipping of the spoil in addition to satisfying the TML criteria for the Crossrail project. While the properties of the spoil can be modified by changing the conditioning treatments, the most suitable properties for TBM operations may differ from those required to optimise the spoil transport and disposal process.

The clogging potential or ‘stickiness’ of London Clay can be considered as an example. Figure 5 shows that in‐situ London Clay has a high clogging potential based on the classification scheme of Thewes and Burger (2005). Reducing the clogging potential can improve the TBM operation and the efficiency of spoil transport and handling processes. The conditioning used for London Clay during tunnelling on Drive X increased the water content and reduced the consistency index slightly from the in‐situ range. This gave spoil strengths of about 20 to 40 kPa which are suitable for shipping, but the clogging potential remains generally high (see Figure 5). From the laboratory index tests, conditioning treatments that increased the water content of London Clay by about 20% typically reduced the shear strength to less than 10 kPa. This conditioning would result in a low clogging potential (see Figure 5), but the strength would be too low to meet the project‐specific shipping criteria.

This example illustrates how the spoil properties required to optimise different stages of the tunnelling and spoil disposal processes can differ. The overall construction process should be considered during project planning and testing and selection of soil conditioning treatments.

5   Port of Miami Tunnel, Miami, Florida, USA

The Port of Miami Tunnel project involved construction of 1.2km long, twin‐bore 11.3m diameter tunnels through mixed limestone and sands and highly porous coralline limestone with high groundwater pressures and permeability. The tunnels were constructed using a hybrid TBM operated in EPB mode in ground conditions where the excavated material could be conditioned to form a stable soil paste.

Due to the highly variable and challenging ground conditions, it was necessary to investigate the feasibility of effectively conditioning the different soil/rock materials to be encountered. A laboratory index test programme was performed prior to tunnelling to inform selection of the most appropriate tunnelling method and effective conditioning treatments for the materials suitable for EPB tunnelling. Further details of the project and the pre‐construction soil conditioning laboratory tests are reported in Merritt et al (2013).

Index tests were performed on samples of the different soil/rock layers mixed in proportions representative of the expected tunnel face conditions. The samples were obtained from drilled shafts to model the effects of the tunnelling process on the in‐situ materials. Key issues influencing the samples properties were the grading and water content. The samples were prepared to represent the expected state of the excavated, conditioned material produced by the TBM as realistically as possible. Due to the variability of the in‐situ materials a wide range of soil mixtures, water contents and conditioning treatments were tested.

Slump tests were performed to allow classification of the conditioned sand / gravel samples as ‘Suitable’, ‘Borderline’ or ‘Not Suitable’ according to the scheme of Peila et al (2009). Vane shear strengths were measured to relate to the sample consistency indicated by the slump tests. Sample gradings were also determined to aid interpretation of the test results based on the classification scheme for typical conditioning treatments for soils with different gradings and groundwater pressures proposed by Thewes (2007).

The index tests showed that it was feasible to effectively condition most of the soil/rock materials using foams and polymers. However, samples containing significant proportions of the highly porous corralline limestone could not be effectively conditioned to form a paste suitable for controlled EPB tunnelling. Due to the high and variable water content and very coarse grading it was not feasible to achieve suitable conditioned samples even with large volumes of specialised polymers in combination with a thick mortar conditioning agent used to improve the spoil grading.

The index tests also provided insights into the effects of the conditioning treatments on spoil properties. The sample consistency indicated by the slump tests and the associated shear strengths were dependent on the combination of FIR and water content. The polymers aided paste formation and water absorption, acting to stiffen the spoil as indicated by a reduced slump and higher shear strength. The index test results also allowed identification of shear strengths ranges for the different sample classifications based on the slump tests.

Following the index tests, a series of pressurised mixing tests and model EPB screw conveyor tests were performed on conditioned samples of the project soils. These advanced tests confirmed the conclusions from the index tests under more realistic test conditions.

During tunnelling, regular TBM spoil testing was performed using the same index tests and sample characterisation tests. The results showed similar spoil properties and effects of the conditioning treatments as observed in the pre‐construction index tests. The TBM spoil tests aided monitoring of the TBM operations and conditioning treatments, and of the ground conditions encountered during tunnelling against the project ground model.

6   Conclusions

This paper has discussed laboratory testing for soil conditioning in EPB tunnelling projects.

Simple index tests prior to tunnelling aid selection of effective conditioning treatments and inform other aspects of construction planning, such as environmental approvals and advance knowledge of the expected TBM spoil properties for disposal. Advanced test methods can also be used to provide more detailed understanding of conditioned soil properties and the excavation process, particularly for projects in difficult ground conditions.

Testing of TBM spoil during construction is beneficial for monitoring TBM operations and the effects of soil conditioning treatments, as well as monitoring the spoil properties relevant for the project‐ specific transportation and disposal processes for the excavated material.

Examples of soil conditioning tests carried out on some projects in a range of ground conditions were discussed. Different index test methods are used for different types of ground to measure the relevant properties, and simple test methods are able to produce samples with properties representative of those observed from the TBM spoil.

As different stages of the construction process can require the excavated material to have different properties to optimise operations, the overall tunnelling and spoil disposal processes should be considered in the planning and application of soil conditioning.

References

Borghi, F.X. & Mair, R.J. (2008). The effects of soil conditioning on the operation of Earth Pressure Balance machines. World Tunnel Congress 2008, India, p. 1469 – 1478.
Davis, J. & Russell, L. (2014). The Crossrail Project, London. The transport and beneficial re‐use of excavated material. SIG Conference, Verona, Italy, May 8‐11 2014.
DAUB (2010). Recommendations for selecting Tunnel Boring Machines, October 2010.
EFNARC (2005). Specification and guidelines for the use of specialist products for mechanised tunnelling in Soft Ground and Hard Rock. EFNARC, April 2005.
Merritt, A.S., Borghi, F.X. & Mair, R.J. (2003). Conditioning of clay soils for earth pressure balance tunnelling machines. Underground Construction 2003, London, p. 455 – 466.
Merritt, A.S. & Mair, R.J. (2006). Mechanics of tunnelling machine screw conveyors: model tests. Geotechnique 56, No. 9, p. 605 – 615.
Merritt, A.S., Jefferis, S., Storry, R.B. & Brais, L. (2013). Soil conditioning laboratory trials for the Port of Miami Tunnel, Miami, Florida, USA. World Tunnel Congress 2013, Geneva, p. 1328 ‐ 1335.
Peila, D., Oggeri, C. & Borio, L. (2009). Using the slump test to assess behaviour of conditioned soil for EPB tunnelling. Environmental and Engineering Geoscience, Vol. XV, No. 3, p. 167 – 174. Pena, M. (2003). Soil conditioning for sands. Tunnels and Tunnelling International, July 2003.
Thewes, M. & Burger, W. (2005). Clogging of TBM drives in clay – identification and mitigation of risks. In Erdem and Solak (eds) “Underground Space Use: Analyses of the Past and Lessons for the Future”, p. 737 – 742.
Thewes, M. (2007). TBM Tunnelling Challenges – redefining the state of the art. Tunel, Collection of Keynote Lectures, ITA‐AITES World Tunnel Congress, Prague, 2007, p. 15‐21.