Soil conditioning plays a key role in earth pressure balance (EPB) methods. The EPB technique enables the construction of near surface tunnels in bad ground conditions, with minimal surface settlement. Soil conditioners are mixed with the excavated earth, altering the soil properties and turning it into a much easier mixture to manipulate.
Notwithstanding the wide use of soil conditioners, little guidance exists regarding their selection and application for tunnelling in different soils. The most common soil conditioning agents used in the tunnelling industry are foams, polymers and bentonites. To achieve an ideal mixture of the excavated soil with the soil conditioner for the operation of the EPB machine, soil conditioning agents can be used alone or as a mixture of different agents at various concentrations and injection ratios.
Research is being carried out at Oxford University to investigate soil conditioning for EPB machines in sand. Our recent investigation has included work on the characterisation of foaming agents, the conditioning of Thanet sand using different agents (foams and polymers), and direct shear box tests on foam-sand mixtures.
Characterisation of foaming agents
In spite of the vast range of foaming agents on the market, it is relatively difficult to find the most efficient foaming agent for a particular ground condition, even though their basic parameters are given: concentration range, foam expansion ratio (FER) and injection ratio.
An attempt was made to select the right foaming agent for a particular type of sand. Five commercial foaming agents were characterised. The FER and 50% (T50%) drainage times were calculated using the methods specified in Ministry of Defence Standard 42-40/Issue 1.
All the foaming agents were used to generate foam with exactly the same settings on a foam generator at Oxford University. The foaming agents are based on anionic surfactants or glycols, with some containing polymer additives.
Figure 1 shows the FER measured at atmospheric pressure for each foaming agent at different concentrations, and the measured 50% drainage time (T50%).
Even though a broad range of concentrations of the foaming agents is recommended by the suppliers (for a better performance), it seems that lower concentrations might be used without obtaining substantial differences between FER at high concentration and FER at low concentration (see Table 1).
As expected, the chemical composition of the foaming agents influences the FER and drainage time. For instance, A at 2% concentration gave a higher FER than D (a foaming agent which contains a polymer additive) at 4% concentration and E shows the longest drainage time even at its lowest concentration, this being longer than for instance D at 4% concentration.
It was not possible to generate foam with agent B, therefore no results for this foam appear in Figure 1. It did not generate foam because the water used in the mix was too hard. This sends a warning to contractors to identify the hardness of their water supply on site. They should either avoid “hard” water supplies or avoid foaming agents with high sensitivity to the hardness of water.
Index tests for conditioned sand
The optimum performance of EPB machines can be reached when conditioning agents are injected, to be mixed with the excavated soil forming a paste, which has better handling properties than the soil itself.
Because the selection of the correct conditioner and its correct amount depends on the properties of the particular excavated soil, the use of index tests (on site) that can give a quick and valid evaluation of the effectiveness of the conditioner is essential.
Index tests have been carried out on Thanet sand to determine its behaviour once it has been mixed with different soil conditioners.
The Thanet sand material was obtained from the CTRL 220 contract site in London. The properties of the Thanet sand tested are:
Slump tests were carried out for four types of commercial foaming agents (the foaming agents described above, except for B) and four types of commercial polymer solutions: PA and PB supplied by S2, PC supplied by S3 and PD supplied by S4.
The mixtures were made up of Thanet sand at 22% water content and foam from one of the four different foaming agents at 65% and 80% foam injection ratio (FIR).
Slump tests
Figure 2 shows the results of the slump tests on mixtures of Thanet sand and foam using four foaming agents at different concentrations and foam injection ratios (FIRs). In addition, Quebaud’s limits5 of the slump of a suitable mix (excavated soil and conditioner) for the EPBM’s optimum operation are shown in the Figures 2 and 3. The area between the black dashed horizontal lines is the area defined by Quebaud5 as the area of “very plastic” mixtures. In addition, the solid horizontal line represents the slump value for an ideal mixture of soil and soil conditioners5 for the EPBM.
Figure 3 shows the variation in slump for the mixtures of Thanet sand and polymer solutions. The results show the variation of slump at different polymer injection ratios (PIRs) (slump tests at different polymer concentrations were also carried out but these results are not shown here).
Furthermore, water at an injection ratio of 25% was added to the Thanet sand (this was the same volume as used for polymers). The final water content of the mixture was 34.8%. Its slump was 240mm.
Most of the foamed Thanet sand tests showed a “very plastic” slump (between 100mm and 150mm slump), with just C at 1.5% concentration and 80% FIR showing a “fluid” slump (over 160mm, see5).
In conclusion, in order to increase the slump of the mixtures (foam Thanet sand or polymer Thanet sand mixture) it is necessary to increase the FIR or PIR, moving the conditioned soil from a plastic slump (slump between 50mm and 90mm) towards a fluid slump.
When just water is added to Thanet sand, a fluid slump is generated (slump higher than 160mm), which is not optimum for an EPBM, which works better with foamed sand mixtures with slump between 100mm and 150mm (between the dashed lines in Figures 2 and 3). This confirms the advantage of using soil conditioners to obtain suitable properties of excavated soil in an EPBM.
The above evaluation, together with the simplicity of the procedure, support the recommendation of using the slump test on site as a first tool to assess the behaviour of soil when it is mixed with soil conditioner. This will allow contractors to decide, for instance, what FIR or PIR is needed for a particular soil.
Flow cone test
Tests were also carried out on a “flow cone” with a view to using this test too as an index of conditioned soil behaviour. Surprisingly, neither the mixtures of foam and Thanet sand nor polymer and Thanet sand flowed through the flow cone apparatus. Usually, once the mixture was placed in the apparatus it dried of excess liquid, and so did not flow continuously. After about one second, the remaining mixture plugged the nozzle, preventing further flow. A higher amount of liquid content would be necessary to make the foamed sand mixtures run continuously out of the flow cone. That type of mixture is not suitable for EPBM operations, because it will have a high permeability, creating an undesirable flow of water out of the screw conveyor, decreasing pressure and making the flow of material difficult to control.
Direct shear box tests
A series of ten shear box tests was carried out on Thanet sand mixed with four different foaming agents and two shear box tests on saturated Thanet sand with different voids ratios. The concentrations of the foaming agents were the same as those used in the index tests.
A direct shear box approximately 250x150x150mm was used to carry out the tests. Measurement of pore pressures, and thus calculation of effective stress, was possible with the apparatus.
Figure 4 shows the results for one of the shear box tests carried out on foamed Thanet sand. In this particular case, the test shows a mixture of foaming agent D with Thanet sand (FIR% of 47.6 and voids ratio of 1.27). The average vertical stress value is 46kPa.
The use of foam increases the voids ratio of the sand to values higher than those reached by sand alone. For example, in Figure 4 the voids ratio of the mixture is 1.27. The results are analysed in the context of the relative density (Dr) of the sand, which provides a good correlation with the shear strength1. The relative density (Dr) defines the relationship between the voids ratio e and the limiting voids ratios emax and emin (the maximum and minimum voids ratios normally attainable for a given sand). A relative density value of 0 means the loosest state of the soil and a relative value of 1 means the densest state. When foam is added to sand, however, negative relative densities (not normally attainable) can be obtained.
The shear stress obtained in a direct shear test is reduced by the use of foaming agents. This is observed in Figure 5, where all twelve tests are plotted. The use of foaming agents can reduce the shear stress by at least 25% compared with saturated Thanet sand tests. However, the high pore pressure values leads to the conclusion that this decrease in shear stress is due to the reduction of effective stress in the shear box tests on foamed sand. This is shown in Figure 6, where the variation of the pore water pressure/vertical stress ratio with relative density is plotted. It is clear that the tests which show higher values of pore pressure also show lower values in shear stress. Taking into account the pore water pressure it is possible to see the variation of the angle of friction with the relative density, as shown in Figure 7.
Bolton’s correlation1,2 is shown on the figure, assuming a critical state angle of shearing resistance of 34º (a typical critical state angle of shearing resistance for a quartz sand). This correlation is used because it addresses the influence of the voids ratio on the angle of friction.
A previous treatment of direct shear box test results4 with foamed soils did not consider the pore water pressure in the mixes during the tests. However, taking into account pore water pressure, it can be observed that the angle of friction for foamed Thanet sand is not much lower than the critical state angle of friction for Thanet sand. It is the high pore pressures that reduce the strength, indicating that permeability characteristics will be important.
The low shear strength of the foam-soil mixture is particularly significant for the tunnelling industry. A lower strength has implications for reduced power consumption and reduced wear throughout the tunnelling process.
An attempt was made to measure the drainage of liquid from the shear band in the shear box. It was observed that for different foaming agents the percentage of liquid drained from the shear band is different. This is seen in Figure 8 where, surprisingly, the highest percentage of liquid drained from the shear band was measured for foam E. Looking at the stability test results in Figure 1, E was expected to have a more stable foam and stay longer with the sand. This reveals that there are other parameters to be considered in the stability of the foam-sand mixture. The mechanisms are not yet fully understood, but the permeability of the soil is expected to play an important role.
Finally, it is interesting to relate the results obtained from the shear box tests with the slump tests. Figure 9 shows the analysis for E (1.5% concentration) at two different FIRs. Here it seen how a high value in the slump test is related to a low value of shear strength and vice versa.
Conclusions
Foam can be used to increase the voids ratio of the soil. At high voids ratios the strength of foam sand mixtures is much lower than at “conventional” voids ratios for sand. This appears to be because of pore water pressure developed in the foam sand mixtures, which until now was not taken into account in the shear stress analysis made for foam-sand mixtures3,4. If pore pressure controls the strength of the foam-sand mixture, it is very important to understand the flow of the liquid through the sand; and so a permeability test is needed to complement and enhance these results.
The angle of friction for foamed Thanet sand mixtures is not considerably lower than the critical state angle of shearing resistance for Thanet sand alone: this differs from the earlier conclusion by Psomas in 20014.
It is clear that the flow cone test is not suitable to determine the fluidity of foam Thanet sand mixtures which are used in the tunnelling industry. The slump test, however, provides a simple workability index for foam sand mixtures and offers a simple and rapid indication of strength for foam sand mixtures (see Figure 9).
Related Files
Figure 8
Figure 2 – Slump test results of foaming agents at different FIR values
Fig 5 – Variation of shear stress with the relative density at 15mm of horizontal displacement in shear box test
Fig 6 – Variation of pore water pressure with relative density at 15mm of horizontal displacement in shear box test
Fig 7 – Variation of angle of friction with relative density at 15mm of horizontal displacement in shear box test
Figure 9
Figure 3 – Slump values of mixtures of Thanet sand with Polymer solutions (at different concentrations and PIRs)
Figure 4
Figure 1 – Foam expansion ratio and liquid drainage time at different concentrations of foaming agents
