Maximum jacking forces of EPB shield tunnelling machine in soft soil

S. Rodenhuis Ministry of Transport, Public Works and Water Management A.M. Talmon Delft Hydraulics

ABSTRACT: The Botlek Railway Tunnel in The Netherlands is a pilot project carried out by a large diameter tunnelling machine with EPB shield, which has been successfully completed. The alignment of this tunnel passes through soft soil types varying from clay with peat to coarse sand, nearing shingle. During an incident at the Northern tube of the Botlek Railway Tunnel the tunnelling process was stopped for a period of nearly one-month. After restarting the tunnelling machine extreme jacking forces were measured. The relation is distinguished between time for standstill, speed of progress and resistance to progress for a tunnelling machine in EPB mode, in the framework of the type of soil and pressures present at the considered location. It is shown that in clayey soil the effect of overcutters is only limited and extreme axial forces can only be avoided by maintaining a drilling speed of more than 20 mm/min. 1 INTRODUCTION In The Netherlands two pilot projects have been carried out concerning shield tunnelling in soft soil. The main objective of these pilot projects is to obtain more experien ce with this construction technique in typical Dutch soil conditions. In an extensive part of the Netherlands the soil conditions are characterised by a high phreatic surface and a very low stiffness. Both projects, the Second Heinenoord Tunnel with a hydroshield and the Botlek Railway Tunnel with an EPB shield have been successfully completed. 1.1 Research program at the Botlek Railway Tunnel This research program is executed under supervision of the Centre of Underground Construction (COB). In this seco nd pilot project, the COB research program focuses on the following items: tunnelling technology, grouting process and control, geotechnical monitoring, construction of cross passages, reusability of outcoming soil, static behaviour of the tunnel lining and effects of vibrations on tunnel and surroundings. The research program has been subdivided into separate areas embodied in different research committees. The research presented here is emphasised on the tunnelling machine itself. 1.2 Tunnel location The Botlek Railway Tunnel is the first Dutch shield driven tunnel for the purpose of railway i n f r a s t r u c t u r e a n d i s p a r t o f t h e s o-called "Betuweroute", which will connect the Port of Rotterdam with the German "Ruhrgebiet". The tunnel has been constructed by BTC Botlek vof comprising of Ballast Nedam Infra, HBG Civiel, Strukton Betonbouw, Van Hattum & Blankevoort and Wayss & Freytag. It crosses the Oude Maas River and it consists of two tubes with an external diameter of 9.45m (concrete lining) and internal diameter of 8.65m, the length is 1850m. Typical soil conditions over the stretch of the tunnel are clay to sandy, on some locations nearing shingle. 1.3 TBM (tunnelling machine) The TBM is shown in Figure 1. It has a cutting wheel with a diameter 9.755m. The soil removed by the cutting wheel is collected and prepared for transportation in the pressure chamber. The length of the shield of the TBM from face to tail is about 9m. The cutting wheel is also supplied with rotating and stator arms in the pressure chamber. Together these arms provide mixing of the removed soil. A hydraulically driven screw conveyor transports the soil from this pressure chamber towards the aft side of the machine and

normally provides for reduction of soil pressure to atmospheric value. A sliding valve, which may be used as a throttle and two piston-type pumps, can also be used for pressure reduction and regulating transportation. This assistance may be necessary in the case of more fluid media. On the aft side of the TBM regular systems provide the further transportation of soil in either dry or fluid state. Preparation of the soil in the pressure chamber prior to transportation is achieved by injection of foam and mixing due to cutting wheel rotation, taking place at the cutting wheel and its rotating arms and stators in the pressure chamber. The foam consists of fluid and a high percentage of air. It's objectives are unbinding the soil particles, regulation of pressure in the pressure chamber and (partially) closing the pores of the face. The cutting wheel can be moved in axial direction by means of hydraulic cylinders, thus allowing for the possibility to measure the total axial force on the cutting wheel. On the outer diameter of the cutting wheel overcutters are provided. With these teeth it is possible to increase the effective diameter of the cutting wheel, in which case free space for the TBM is increased. The objectives for the use of these "overcutters" are reduction of friction on the TBM-shell and allowance for sharper curves (steering). Tunnelling process is intermittent; the TBM is stopped during construction of each ring of the tunnel. Progress of the TBM between construction of subsequent rings is approximately 1.5m.

considerable amount of clay. Peaks in jacking forces are found on other locations as well, nearing but never reaching this level. In order to compress the area of interest here the focus is laid on the measurements around this incident. 2 MEASUREMENTS AND EVALUATION 2.1 Outline of research of measured data In order to understand the nature of the extreme jacking forces a comparison will be made to the data measured on locations where an undisturbed tunnelling process was found before and after the standstill. The data will be reviewed in the framework of comparable soil and pressure circumstances. Therefore the last meters of prog r e s s p r e c e d i n g t h e s t a n d s t i l l a n d t h e following meters have been investigated in detail. Available data have been registered at intervals of approximately 5 seconds. Data of deformation on the surface level above the tunnel were not available for this ass essment. 2.2 Data assignment All data are assigned to identification of the ring next for construction; this refers for instance to ring numbers stated in the graphs and samples, which were taken. The actual face-position is about 7m away from the stated ring due to the length of the TBM. 2.3 Parameters and influences The axial force balance of the TBM is described in Figure 2. The symbols are defined below: Fj : Total jacking force Fcw : Total axial force on cutting wheel p ch : A v e r a g e s o i l p r e s s u r e i n p ressure chamber Ffr : Friction force.
Ffr p tail Fj TBM Fj p ch Fcw

Figure 1; TBM overview.

1.4 Incident during tunnelling I n t he progress of the TBM preceding the construction of ring 302 of the Northern tube mechanical damage to the propulsion of the cutting wheel was detected. After investigating and testing extensive repairs had to take place. The total time of this standstill has been about 26 days. After restart of the TBM extreme jacking forces were measured. On this location the soil is sandy with a

p tail Ff r

Figure 2; Axial force balance of the TBM.

The soil pressure at the tail gap is not measured.

A correction for the resulting force is made by deduction of the tail gap frontal area from the frontal area of the pressure chamber (thus assuming equal pressures in the pressure chamber and tail gap). The centre plate of the cutting wheel (covering the propulsion and hydraulic cylinders) is deducted from the frontal area of the pressure chamber, as the pressure on this part is included in the measured axial force on the cutting wheel. The TBM also pulls trailing lorries with machinery. The force required for the lorries is included in the friction force. The equation of axial force balance is shown in equation 1: Fj = p ch * A ch + Fcw + Ff (eq. 1) with A ch : Effective frontal area on which p ch acts. Normally friction force is a given fact, not directly influenced by the speed of the TBM. This implies that the balance is obtained by the adjustment of Fcw a n d p ch t o Fj . B o t h F cw a n d p ch a r e components, which can directly be influenced by the way the TBM is operated. Friction force on TBM F fr is the sum of friction between the surrounding soil and the shield, friction between the seals at the tail and the tunnel tube and towing resistance of the trailing lorries (the latter is measured and found insignificant). The axial component of the gravity force on a sloping stretch is not taken into account; this component is insignificant due to limited slope of this railway tunnel. With the measured data it is possible to calculate the total friction on the TBM. Similarly, but not shown, it is possible to investigate the moments with respect to the transversal and vertical axis, of which the balance indicates effects of steering on pressure on the shield and the seals at the tail of the TBM. Influences on components in equation 1: Average pressure p ch in the pressure chamber is influenced by the balance of soil input and transportation by the screw conveyor and, dependent on the porosity of the face, the foam injection. Soil input is primarily dependent on the speed of the TBM, the transportation outward is indicated by the measurements of the screw conveyor. Total axial force on the cutting wheel F cw is the sum of pressure difference between the face and pressure chamber, acting on the surface of the wheel (open sections extracted) and the pressure in the chamber on the central cover of the wheel (covering propulsion etc.) and total axial force on cutting teeth.

Friction on the shield is thought to be dependent of the effective stress and possibly also of active steering forces. Only p ch and F cw are directly influenced by the operator within less than a few minutes. 3 MEASURED AXIAL FORCES The incident takes place during the drilling for ring 302, drilling for this ring was not completed when the incident occurred. Figures 3 and 4 show graphs for distinct rings with measured and calculated values of forces, speed and pitching versus time. T h e c u r v e s f o r F j (F_j) as represented are calculated based on oil pressures in the jacking system. No efficiency losses have been deducted. The efficiency losses used to calculate Ffr (F_fr) through equation 1 are assumed at 10% of F j .
Ring 0301
50000 F_j F_fr F_cw p_ch * A_ch

40000

Force [kN]

30000

20000

10000

Speed TBM

[mm/min]

60 40 20 0 -240

[mm/100m]

Pitch TBM

-250 -260 00:20 00:30 00:40 Time [hh:mm] 00:50 01:00 01:10

Ring 0302a
50000

F_j F_fr

F_cw p_ch * A_ch

40000 30000

Force [kN]

20000

10000 0

Speed TBM

60

[mm/min]

40 20 0 -240 -250 -260 01:50 02:00 02:10 02:20 02:30 Time [hh:mm] 02:40 02:50 03:00

[mm/100m]

Pitch TBM

Ring 0303
60000 50000 40000

F_j F_fr

F_cw p_ch * A_ch

Ring 0310
F_j 50000 F_fr F_cw p_ch * A_ch

40000

Force [kN]

30000 20000

Force [kN]

30000

20000

10000 10000 60

Speed TBM

60

Speed TBM

[mm/min]

[mm/min]

40 20 0

40 20 0 -270

-240

[mm/100m]

Pitch TBM

[mm/100m]

-280 -290 13:00 13:30 14:00 Time [hh:mm] 14:30 15:00

Pitch TBM

-250 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 24:00

Time [hh:mm]

Figure 4; Forces, speed and pitching versus time.

3.1 Evaluation The effect of the 26 days standstill is clearly shown in the friction force F_fr for ring 303 and decreases slowly over the subsequent rings. The stretch after the stand still needed for the decrease of the effects of the standstill until normal conditions are measured again is found to be relatively long, as ring 305 still shows friction forces with values about 1.5 times the values of ring 301. At the progress for ring 307 the friction has decreased significantly compared to the values for ring 303. Now most of the 9m long shield of the TBM has passed the area where the TBM was stopped during the standstill. Variations in pitching of the TBM indicate that the machine floats up due to buoyancy and variation in the amount of outcoming soil in the pressure chamber when at standstill or at very low speed. During the 26 days standstill in the alignment for ring 0302 this displacement was limited to 1 mm at the cutting wheel, when assuming the tail of the machine as fixed in vertical position by the tunnel tube. During progress pitching variations are related to speed (with some delay), but they are not synchronically to the variations in calculated friction. Pitching variations and steering during progress have been found not to affect the total friction force, a conclusion that is assumed to be related to the use of overcutters at the subject stretch. Note the synchronicity of speed variations with the friction F_fr. An increase of speed leads to a decrease in friction. The variation of friction follows the speed within a couple of seconds. This effect is clearly visible for ring 307 but also noticeable for the previous rings. Especially ring 0310 shows the time-delay in the relation between variation in friction and speed. This turns out to be systematic for soils with a significant

40000

Force [kN]

30000

20000 10000

Speed TBM

80 60 40 20 0

[mm/min] [mm/100m]

Pitch TBM

-250 -260 07:50 08:00 08:10 08:20 08:30 08:40 08:50 09:00

Time [hh:mm]

Figure 3; Forces, speed and pitching versus time.

40000

Force [kN]

30000 20000

10000

Speed TBM

80 60 40 20 0

[mm/min] [mm/100m]

-250

Pitch TBM

-260 12:40 12:50 13:00 13:10 13:20 13:30 13:40 13:50 14:00

Time [hh:mm]

Ring 0307
F_j 50000 F_fr F_cw p_ch * A_ch

40000

Force [kN]

30000

20000

10000

Speed TBM

60

[mm/min]

40 20 0 -250

[mm/100m]

Pitch TBM

-260 -270 19:30 20:00 20:30 21:00 21:30

Time [hh:mm]

percentage of clay, and leads to decrease of friction where the speed is over 20mm/min and an increase of friction where the speed decreases below 10mm/min. In between these values is a transition area. Well above 20mm/min the friction decreases faster but the final value seems not to differ from the one found at speeds of about 20 mm/min. The stabilised values for friction for speeds over 20mm/min are approximately 65% of the extreme values at low speeds. However, long standstill periods may lead to an increase of 150% compared to the values of continuous progress at speeds of 20mm/min or more, as shown when comparing the data of rings 301 and 302 to 303. Although the overcutters have been used, no media were injected behind the cutting wheel to slow down deformation of the surrounding soil towards the shell of the TBM. Measured data are not available to determine whether it is the surrounding s oil which reaches the shell or the grout injected at the tail gap which proceeds forward in the gap between soil and shell. In view of the slow decrease of friction assumed is that soil deformation is decisive. Measured data of the further stretch differ more and more, obviously influenced by a change in soil composition (more sandy, according to the geotechnical profile), as the graph for ring 0310 indicates. Friction decreases for normal progress and eventually the overcutters are switched off.

the cutting wheel. 4 GEOTECHNICAL PROPERTIES Granular division is measured for ring 310 (Grote, 2001), while the face is at position of ring 0315. Four samples are taken from the pressure chamber, representing the mixture of the removed soil. The measurements are shown in Table 2. Note that the shield of the TBM is in contact with soil, which is not mixed but layered, some layers sandy and some layers clayey. At ring 0301 the amount of clay is higher according to the geotechnical profile, therefor the measurements shown in table 2 are only indicative for the soil around ring 0301. The portion of clay decreases gradually with progress towards the values of ring 0310. Comments from visual inspection of the samples are: Plastic mixture Lot of foam Lumps of clay
Grain diameter d Cumulative weight [mm] of grains < d [%] 0.002 5 0.005 8 0.01 9 0.02 10.5 0.05 13 0.08 20 0.1 25 0.14 50 0.2 67 0.26 80 0.5 93 1 98 2 99.5 Table 2; Granular division of ring 0310. Type of soil Fine silt Normal silt Coarse silt Fine sand Fine sand Normal sand Coarse sand Fine shingle

3.2 Summary For the locations with extreme (0303) and normal (0301 and 0302) jacking forces the simultaneous values of components in the equation of axial balance are stated in table 1, valid for the subject stretch.
Fj F cw pch * Ach [kN] [kN] [kN] Extreme 55,000 6000 14,500 Normal 41,000 8000 15,000 Table 1; Normal and extreme forces axial on TBM. F fr [kN] 31,000 12,500

For data of ring 0303 the centre of the TBM is located around ring 304, the data on vertical pressures and stresses on the present location are shown in Table 3. These vertical pressures are derived from ring 0320 (Talmon & Bezuijen, 1999) and corrected for actual depth position of ring 0304.
Ring 0304 Depth to Total surface stress [m] [kPa] Tunnel top -16.7 313 Centre line -21.45 409 Tunnel bottom -26.2 505 Table 3; Vertical pressure and stress. Waterpress. [kPa] 191 238 285 ' Eff. stress. [kPa] 122 171 220

It is not a coincidence that both F cw and p ch are found to be reduced at the time the extreme jacking forces are found: This is a combined res ult of low speed of progress and attempts by the operator to keep the TBM in motion by reducing the backwards -directed forces. The latter is merely achieved by increasing the ratio of speed of the screw conveyor over soil input. The slow speed reduces the axial force on

The effective stress ' is in fact the vertical grain stress. The soil loading on the TBM is a combination of horizontal and vertical effective

stresses plus the water pressure. 5 FRICTION AND EFFECTIVE STRESS In this case calculations will be kept very brief and simple. Total surface area of the TBM shield is: A shield = * 9.755 * 9 = 276 m. Average friction stress fr will be related to vertical effective stress, resulting in a friction coefficient f, calculated through equations 2 and 3, results are shown in Table 4. fr = Ffr / A f = fr / ' (eq. 2) (eq. 3)

This means that a high drilling speed is necessary when high axial forces are encountered. This is against the practice noticed at the Botlek Railway tunnel to reduce drilling speed in such situations. 7 ACKNOWLEDGEMENTS The research data presented here are derived from the COB-pilot project F300 in which the authors participate; this research project has not been finished yet at this stage. The authors are very grateful to both COB and the builder of the tunnel BTC Botlek vof, with the granted opportunity to publish these data. The assistance of J.F.W. Joustra, MSc, B.J.H. Grote, MSc and A. Bezuijen, MSc has been highly helpful. REFERENCES Talmon, A.M. & Bezuijen, A., 1999 C980, K 3 0 0 Pr e d i c t i e s s t a b i l i t e i t e n drukverdeling, Grote, B.J.H. 2001 K300-W-12-4, TBM Botlekspoortunnel; Meetrapport TBM metingen, deel 4 Noordelijke buis,. Heij, R.A. de & Kessel, A.J. van, Wauben, M . 1 9 9 9 K 3 0 0-W-013, Predicties Krachtenbalans, Groutproces en Slijtage snijtanden Botlekspoortunnel, Rodenhuis, S. (in progress).F300-W-01804, Botlekspoortunnel, Evaluatiefase A, deel 4 Noordelijke buis MQ2, 4 en 5, Noordelijke en Zuidelijke buis : axiale krachtenbalans,

Extreme Normal fr 112.3 kPa 45.3 kPa f 0.656 0.265 Table 4; Average friction stress and friction coefficient.

There are different calculation methods in use for soil loading on TBMs. These involve different types of assumptions. The f-value as calculated here serves to indicate the order of magnitude of soil friction, and is an aid to compare the starting processes with stationary conditions. The order of magnitude of the friction coefficient under stationary conditions agrees with know values. The origin of the high friction at start needs to be established. It is expected to be due to relaxation of the soil during standstill. 6 CONCLUSION Extreme jacking forces for tunnelling machines in EPB mode in soils with a significant percentage of clay are caused by extreme friction. Due to reduced speed of progress and actions by the operator the axial force on cutting wheel and pressure in the pressure chamber usually are reduced when extreme jacking forces occur, but the total reduction obtained by these actions is relatively small. The incident, which was investigated, shows the limited effect of the use of overcutters on friction forces after a long standstill period in soils with significant percentage of clay, even if these overcutters were already used prior to the standstill. Friction may be reduced in these soils by shortening the time for stands till and keeping up the drilling speed of progress well above 20 mm/min.