Characterization of soil conditioning for mechanized tunnelling

May 07, 2008

The EPB tunnelling has achieved a higher number of application and this trend is clear both for large size machines and for small diameter machines. In order to extend the application field for EBP machines, the natural soil has to be treated with additives during the excavation. The results of a test program on different conditioned not-cohesive soils by means of slump cone test are presented in the paper. The water content and the amount of conditioning foam were varied in the tests in order to find the influence on the quality of the conditioned ground and how the foam and water contents influence the mix behaviour.

Earth Pressure Balance shield tunnelling has been successfully adopted in urban environments in recent years for very different ground conditions and at present it can be considered the most commons used mechanized tunnelling technology. This increase in number of tunnels excavated by EPB machines has been possible, due to the mechanical and electronic improvements and to the more effective use of additives injected into the soil and in the bulk chamber to condition the soil.
For EPB application the soil must have defined properties: in fact, if it is too permeable, groundwater cannot be controlled by the counter-pressure, even if the excavation chamber and the screw conveyor are completely full; if the material is not plastic there is no smooth flow to the screw conveyor, since the material tends to arch at the entrance of the screw conveyor and it is difficult to control the pressure inside the excavation chamber, therefore the stability of the face (Anagnostou and Kovari, 1996) cannot be guaranteed (Figure 1). Moreover, in case of sticky clayey soils there is the risk that the cutter head, the excavation chamber and the screw conveyor can clog up.
As a first assessment, the soil features required for good EPB operations are:
  1. a certain degree of plasticity that makes the treated soil suitable for the pressure transmission in the excavation chamber and the pressure decay control along the screw conveyor, as well as for the controlled extraction through the screw conveyor itself;
  2. a low inner friction of the bulk material that allows to reduce both the power requirement of the cutting head, as well as the wearing of the machine parts that are in contact with the soil;
  3. the persistence of the above-mentioned characteristics over the time, to allow a safe control of the face stability during the whole excavation step and when the machine has to be stopped due to any reason.
In order to extend the original application field for EPB shields towards more difficult soils, as cohesionless material (sand and gravel) or sticking clayey soils, some additives, like foam, polymers or fillers, can be injected and mixed with the ground during the excavation and extraction step, improving some features of the soil like plasticity.. It is worth to point out that the natural workability of the soil, i.e. without additives injection, is influenced mainly by the finer content and the water content.
For a correct design development and work control during EPB operation it is necessary to develop a specific series of tests able to give clear indications of the conditioned material behaviour.
The technical literature regarding test methods for the characterization of conditioned soil is limited and for these researches tests derived from the geotechnical or concrete measurement technology have been used: mixing test; cone penetration test; permeability test; compressibility test; shear test; slump cone test (Milligan 2000,2001; EFNARC 2005). Some large scale tests using a laboratory screw conveyor device have been recently proposed and this new procedure has shown its feasibility.
Slump cone test has been widely used to provide a measure of the conditioned soil plasticity by Peron & Marcheselli (1994), Maidl (1995), Quebaud et al. (1998), Bordachar and Nicolas (1998), Jancsecz et al. (1999), Williamson et al. (1999), Langmaack (2000) and Vinai et al. (2007) since is simple to be carried out, quick and low cost and gives an overall index of the conditioned material behaviour. These authors suggested that good values of slump cone fall range from 120mm to 250 mm.
Peña (2004) using the slump test focused his research mainly on the characterisation of foaming agents to condition a reference sand and he compared five different foaming agents and two polymers. He noticed that for the tested sand, to obtain a plastic behaviour and a slump cone fall of 100-150 mm, it was necessary to have a water content of 22%,a concentration of foaming agent around 2% and a FIR between 65% and 75%.
Anyway only a test that is able to simulate EPB operations and which involves both the conditioning of the soil and the interaction between the soil and the machine would be truly significant to understand the conditioned soil behaviour as done in the following researches.
  • a 1/10 scaled model for EPB excavation simulation was set up in the work of a National French Research Programme (AFTES, 2001). The model consisted of a 500mm diameter cutting head of 500 mm, a conical excavation chamber, a inclined screw conveyor, a horizontal screw conveyor, a cylindrical shield and four thrusting cylinders. The soil to be excavated was placed in a rigid 2 x 1.3 x 1.3 m box. A loading device, using ten air cushions, allowed the simulation of an additional overburden. The system was instrumented with several monitoring transducers for the driving parameters, the soil stresses and deformation control (Branque et al., 2003);
  • a laboratory model of a screw conveyor device and a full-scale EPB machine screw conveyor were used by Bezuijen and Schaminée (2001) to study the behaviour of conditioned sand soils. They observed that the pressure was dissipated linearly along the conveyor, that the screw torque was approximately constant and that the pressure at the end of the conveyor was depending on the opening of the gate valve at the discharge point;
  • a full-scale EPB screw conveyor was used by Yoshikawa (1996) and a number of tests were performed with plastic soil and with different screw speed. He observed that a linear pressure gradient was present in the screw conveyor;
  • a laboratory screw conveyor apparatus, where the material was extracted from a tank by a sub-horizontal screw, was built by Oxford University, in partnership with Cambridge University (Peña, 2003; Mair et al., 2003; Merritt and Mair, 2006). The laboratory device was made up of a pressurized tank which was connected to a 1m long and 0.1m diameter horizontal screw conveyor. The screw conveyor was instrumented in four sections, each with two load cells to measure the total normal stress and the soil-casing interface shear stress components and a pressure transducer to measure the pore water pressure in the soil. Measurements of the screw torque were also carried out. Tests can be performed with various pressures applied to a cohesive soil over a range of screw speeds, with different discharge outlet conditions.
  • tank by an inclined screw, was built in Politecnico di Torino (Vinai et al., 2006). The device represent an approximate 1:10 screw conveyor scale model with a screw conveyor lenght/diameter of 9. The device (Figure 2) is made of a 800mm high tank with a 600mm nominal diameter which is filled with ground. An alluminium plate connected to a hydraulic jack, with a stroke of 500mm, applies a nominal pressure to the tank (up to 2MPa). A 1500 mm long screw conveyor is coupled to the tank with an upward inclination of 30° and the screw extends inside the tank to collect and extract the soil. The diameter of the screw case is 168 mm, the flights have a pitch of 100 mm.
During the test the following date are monitored: the pressure distribution in the tank, the required torque of the screw, the upper plate displacement gradient, the discharged weight gradient and the trend of the pressure values along the screw conveyor. These parameters allow a good characterization of the conditioned soil.
To verify the applicability and feasibility of slump test to characterize conditioned soil, some soil sets mixed with different percentages of water and foam were tested. The used foam is obtained with a FER (ratio between the obtained volume of foam and the volume of the generator fluid: water + foaming agent) of 16, which is an average value with reference of the one usually used in tunnelling, while the foaming agent is a commercial one (Polyfoamer used with a concentration of 2%). Since in practice, in real tunnel excavation, a great variability of FIR can be observed raging from 10-80% (EFNARC 2005), the test are carried out using a FIR range of 20-60%.
The used soils are prepared in laboratory (Figure 3) and their general properties are:
  • soil 1: a medium size sand, with D10 = 0,12 mm and D60 = 0,5 mm
  • soil 2: a mix with the same sand of soil 1 and gravel with grain size of 4-8 mm, with D10 = 0,2 mm and D60 = 5 mm
  • soil 3: a mix with the same sand of soil 1 and gravel with grain size of 8-15 mm, with D10 = 0,2 mm and D60 = 9 mm
  • soil 4: a mix with 45% of sand, 45 % gravel 4-8 mm, 10% of silt, with D10 = 0,2 mm and D60 = 3,5 mm 
3.1 Slump test procedure
The slump cone test is performed following the Standard Test Method for Slump as suggested by ASTM 143C and it is carried out as follows: the conditioning soil is mixed with the foreseen amount of foam and water in a concrete mixer and then it is pound inside 2 slump cones. After one minute, without stroking or mixing the soil the cone is lifted up. The fall value and the global behaviour of the mix is then observed and it is classified using Figure 4 reference schemes.
In the definition of the material behaviour the shape of the slump, the breaking way of the soil cone and the drainage of water from the conditioned soil are observed and taken into account.
Three main behaviour fields are identified:
  • too stiff behaviour or impossibility to create a plastic “paste” ,due to insufficient water or foam content: an irregular collapse of the cone can be observed;
  • too fluid behaviour due to excessive water or foam content: a drainage of water from the soil mass can be observed
  • correct behaviour of the mix: the conditioned soil behaves plastically.
Based on these parameters the mix is then defined as “not suitable”, “borderline” or “suitable”.
The influence of the large size grains on the conditioning procedure is relevant and it’s clearly focused on Figure 5: the reduction of the suitable area induced by a certain amount of gravel (soils 2 and 3) inside sand (soil 1) is evident.
It is therefore relevant to note that it is the sand that interacts with the foam to create a “pulpy paste” encompassing only larger grains which, if they are too many, break the conditioned mass and do not allow to get a plastic paste to be obtained.
Furthermore the presence of silt in the mix makes an enlargement of the suitable area possible since the fine granulometric fraction fills the vacuum between the larger grains, interacting with the water thus permitting less foam to be used.
Soil properties and behaviour have to be changed in order to properly use EPB machines for tunnelling in cohesionless ground. This change is obtained although with injection, inside the bulk chamber and on the cutting wheel, of special additives, such as foam. The definition of the correct amount of conditioning agent and its control during the excavation process are key factors for this mechanized method.
Among the various possibilities, the slump cone test appears to be a simple and low cost procedure that can be used both for job site control and for the preliminary design. In this last case it must be coupled with tests carried out using a screw conveyor device (Vinai et al. 2006; Merritt and Mair 2006).
The carried out test program using slump test on cohesionless soil has allowed an assessment procedure and a reference table to be obtained . The influence of the large size grains on the conditioning effects is evident from the results of the research since it appears a reduction of the suitable area. It is therefore evident that it is the sand that interacts with the foam to create a “pulpy paste” encompassing only larger grains which, if they are too many, break the conditioned mass and do not allow to get a plastic paste to be obtained.
possible since the fine granulometric fraction fills the vacuum between the larger grains, interacting with the water thus permitting less foam to be used. Finally, the carried out tests show that the slump test is a good indicator to define the global behaviour of a conditioned soil and due to its simplicity, can be used in the preliminary design stage but in particular on the job site to keep the conditioning development under control during excavation.
[1] Anagnostou G., Kovari K., 1996, Face stability conditions with Earth-pressure-balanced Shields. Tunnelling and Underground Space Tecnology,11(2), Pergamon Press, Oxford, 165-173.
[2] Milligan G., 2000, Lubrification and soil conditioning in tunnelling pipe jackong and microtunnelling state of the art review, Geotechnical consulting group, London.
[3] Milligan G., 2001, Soil conditioning and lubrification agents in tunnelling and pipe jacking, Proceedings of Undreground Construction 2001, London, 105-116
[4] EFNARC, 2005, “Specification and guidelines for the use of specialist products for Mechanized Tunnelling (TBM) in Soft Ground and Hard Rock,” Recommendation of European Federation of Producers and Contractors of Specialist Products for Structures.
[5] Vinai R., Oggeri C., Peila D., Pelizza S., 2006, “Condizionamento con schiuma dei terreni per applicazioni EPB: sperimentazione mediante un nuovo apparato di laboratorio.” Gallerie e grandi opere sotterranee, 78 (1), pp. 39-47 (in Italian).
[6] Mair R.J., Merritt A.S., Borghi F.X., Yamazaki H. and Minami T., 2003, “Soil conditioning for clay soils,” Tunnels and Tunnelling International, April 2003, pp. 29-32.
[7] Peron J.Y. and Marcheselli P., 1994, “Construction of the 'Passante Ferroviario' link in Milan. Italy. lots 3P, 5P, and 6P: excavation by large EPBS with chemical foam injection,” Proc.,Tunnelling '94, IMM, London, pp. 679 – 707.
[8] Maidl U., 1995, „Erweiterung der Einsatzbereiche der Erddruckschilde durch Bodenkonditionierung mit Schaum,“ PhD Dissertation, Ruhr-Universität Bochum, Germany (in German).
[9] Quebaud S., Sibai M., Henry J.P., 1998, “Use of chemical foam for improvements in drilling by earth pressure balanced shields in granular soils”, Tunnelling and Underground Space Technology, 13(2), pp. 73 – 180.
[10] Bordachar F., Nicolas L., 1998, »Fluides conditionneurs pour la pression de terre », Tunnels et ouvrages souterrains, 169(Janvier/Février), AFTES, pp. 21 – 27.
[11] Jancsecz S., Krause R., Langmaack, L., 1999, “Advantages of soil conditioning in shield tunnelling: experiences of LRTS Izmir.” Proc. International Congress on Challenges for the 21st Century, Alten et al. (eds), Balkema, Rotterdam, pp. 865-875.
[12] Williamson G.E., Traylor, M.T., Higuchi, M., 1999, “Soil conditioning for EPB shield tunneling on the South Bay Ocean Outfall”, In: Proceedings of RETC 1999, pp. 897 – 925.
[13] Langmaack L., 2000, “Advanced technology of Soil Conditioning in EPB Shield Tunnelling,” MBT publication.
[14] Vinai R., Oggeri C., Peila D., 2007, "Soil conditioning of cohesionless sand for EPB applications: a laboratory research", Tunnel and Underground Space Technology (accepted for publication).
[15] Pena M., 2003, “Soil conditioning for sands,” Tunnels and Tunnelling International, July 2003, pp. 40-42.
[16] AFTES, 2001, EUPALINOS 2000, Synthèse, AFTES (ed.),Octobre 2001, Paris
[17] Bezuijen A., Schaminée P.E.L., 2001, Simulation of the EPB-Shield TBM in model tests with foam as additive. In: Proccedings of Congress on Modern tunnelling science and technology, Kyoto, Balkema, Rotterdam, 935-940.
[18] Yoshikawa T., 1996, “Soil pressure drop af the screw conveyor for shielded machines.” Trans. Jpn. Soc. Mech. Engrs, Part C 62 (595), pp. 1197-1203.
[19] Merritt A. and Mair R.J., 2006, “Mechanics of tunnelling machine screw conveyor: model tests.” Geotechnique, 56(9), pp. 605-615.
[20] ASTM C143/C 143M – 00, 2003, Standard test method for Slump of Hydraulic-Cement Concrete, Annual book of ASTM Standards.

1 Politecnico of Turin, Italy
2 CNR-ICAG, Torino, Italy

More News and Articles


Ing. Luca Borio (Dept. of Land, Environment and Geo-technology Politecnico di Torino)

10129 Torino, Italy


+39 011564 7723


+39 011 564 7699



To website