From May to September 2006, the Fira to Park Logistic section of Barcelona’s new metro Line 9 was excavated by a JV of FCC/Ferrovial/OHL/Sacyr. Unlike other portions of Line 9 (T&TI Feb 2005, p26 & T&TI March 2002, p16), where a 10.9m i.d. tunnel is being constructed, this section was excavated with a smaller 9.4m Herrenknecht EPBM installing an 8.32m i.d. segmental lining for a standard twin-track arrangement. The client is GISA, the local government infrastructure agency. PaymaCotas is part of the engineering team in charge of the tunnelling works.
The 1.3km long drive has been used to optimise settlement minimisation techniques for a future section of the Line with very similar soil conditions, which is due to be driven with low overburden under a heavily built-up area of the city. Settlement results are also being used for the evaluation of possible building damage and corresponding ground treatment requirements.
Geology
The 1.3km drive between Fira and Park Logistic stations forms part of the 15km long western ‘Stretch I’ of Line 9, which runs entirely underground. For the first 600m the drive passes under a wide street with little traffic, to later cross the ‘Ronda Litoral’ a principal traffic artery of Barcelona city.
The soils in this area consist of Holocene deposits of the Llobregat River, with very little variation along the alignment. Excavation is predominantly through silty sands with some portions of sandy silts, silts and silty clays. The groundwater level is located 5m below surface (sea level). Overburden is an average of 1.3 diameters.
Tunnel excavation
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalDataExcavation began on 12 May 2006, southwards from Fira station, with a minimum overburden of less than one diameter. With this in mind, and the presence of a main sewer conduit just a few meters above the tunnel crown, an intensive jet grouting programme was carried out to prevent any uncontrolled settlement during the TBM launch.
After the first 140m, where the ground treatment was carried out, and with EPB pressures increasing from 0.7 to 2.1 bar in the crown, as the overburden and water pressure increased, the TBM went on to excavate in untreated ground.
During the drive the TBM achieved a best daily advance of 36m, a best week of 151.5m and a best month of 445m. However, excavation was hindered by some problems, unrelated to the geology or the settlement control, which lowered the overall average to a mere six rings/day. In addition, due to the sand’s abrasiveness, several tools changes were required in hyperbaric conditions. These were carried out in small blocks of treated ground to improve face stability and prevent air blow out due to reduced cover.
Monitoring and data analysis
The main aim of this drive, as indicated earlier, was to find an optimum EPB working method to control volume loss, in order to minimise the requirement for passive ground treatment methods on the future drive. Ground and surface monitoring, to collect all necessary settlement data and corresponding volume losses, as well as record the influence of different EPB excavation parameters, was therefore of the utmost importance.
Surface instrumentation relied on settlement monitoring arrays in the ground and topographic prisms on nearby buildings. The surface monitoring arrays included three ground displacement components, including paired extensometers and inclinometers on each side of the tunnel (reducing the number of boreholes required and assisting data comparison) and an extensometer in the crown. These could be correlated with underground instrumentation, i.e. every inclinometer and extensometer is correlated with the displacement of the instrument head, in order to obtain 3D data records.
The prisms installed on the buildings not only record actual settlements but also angular distortion; this is achieved by placing two or more prisms in a vertical row to provide readings of vertical rotation.
Combining the readings of all the installed instrumentation, it became possible to evaluate the total ground Volume Loss associated with the TBM excavation. Frequency of the readings was very important, to trace the settlement development and correlate it with the shield advance. This is not covered in detail here, but will be the subject of future articles.
The reading frequency for Fira to Park Logistic was developed according to the sequence in Table 1, which starts 100m before and stops 200m behind the passage of the TBM cutterhead, depending on settlement stabilisation.
Settlement studies
Three settlement study zones, clearly differentiated for EPB excavation parameters, were analysed in particular. In the first two zones complete settlement arrays existed. Due to surface access restrictions in the third zone, only instrumentation on the tunnel axis was possible. However, this was still enough to carry out suitable correlations between settlements and volume losses.
In order to evaluate Volume Loss (Vl) the Peck formulation was adopted. The Peck and Schmidt (1969) method for settlement evaluation is based on the hypothesis that the settlement basin generated by the excavation of a tunnel has the shape of a Gaussian curve, which is fundamentally determined by three factors: The excavation geometry (diameter and depth); soil characteristics; and Vl.
The formulation indicates that:
Where: Sv is the settlement at a given horizontal distance (y) from the tunnel axis
Smax is the maximum settlement read on the vertical of the tunnel
i is the inflection point of the Gaussian curve
The i parameter defines the distance between the tunnel’s vertical axis and the inflection point of the Gaussian curve, fixing the boundary between the inner surface compression zone (sagging) and the outer tension zone (hogging).
According to the empirical observations of O’Reilly (1982) and New (1991), it is considered that point i can be related to the tunnel depth Z by the following formulation:
Where: Z is the depth of the tunnel horizontal axis
K is a factor function of the soil, which values can range from 0.3 (granular) to 0.60 (cohesive soils)
According to the properties of the Gaussian curve, it can be estimated that the maximum settlement is a function of the Volume Loss, which is the total area of the curve, based on the following formulation:
Where: Vl is volume loss
And: A is the tunnel section (m2)
Therefore, knowing the shape of the surface settlement basin, as recorded by the installed instrumentation, it is possible to establish the generated volume loss. This task is accomplished by calculating the best fitting Gaussian curve considering the actual tunnel geometry and the recorded settlements and by varying the K and Vl factors.
In this Case study, the variability of factor K (soil dependent) has been considered of little importance since, as previously discussed, the geology along the alignment is fairly uniform.
EPB excavation parameters
Several parameters related to EPB tunnelling are directly linked to the inevitable volume loss generated by excavation. In this case, the following were considered to have the most influence on the phenomena, considering the actual surface and soil conditions:
Face pressure: The basic principle of any EPB excavation is that the applied face pressure should be in equilibrium with soil and water pressures to avoid ‘extrusion’ and consequent volume losses.
In order to minimise ground loss it was necessary to keep the pressure in the crown as close as possible to the design parameters, avoiding unwanted pressure oscillations. It was deemed vital that the density of the material in the chamber was kept as high as possible, in order to have a pressure gradient close to that of the soil at the front. When necessary, bentonite slurry was injected into the chamber during ring build to counteract any drop in pressure.
Shield slurry injection: In granular sandy soils with poor cohesion below the water table, volume losses generated along the shield length can have an important effect on the total Vl (in cohesive soils the importance is less significant).
EPB shields are generally conical, therefore in many cases a lack of immediate support generates an inevitable closure of soil against the steel mantle with resulting surface settlements. Controlled injection of a viscous fluid through the shield has a ‘compensating’ effect.
Tail shield grouting: As the EPBM advanced, the annular void was promptly filled with grout; this is the only effective way to keep settlements low in highly frictional, uncohesive ground. Volume loss largely depends on the efficiency of injection, its pressure and distribution around the segmental lining.
TBM parameters and settlements
At the beginning of the drive, just out of the previously treated ground, large settlements in the range of a couple of inches were recorded. This was certainly due to lack of experience and highlighted the need for tight control of all the TBM parameters. This process required input from all parties (engineer and the contractor, with support from the TBM manufacturer), resulting in a systematic reduction of the volume loss from 1.4% down to 0.45%.
The settlement recorded in Zone 1 can be seen in figure 1 (p42), the correlation between recorded settlements (orange dots) and the Gaussian curve is noteworthy.
In Zone 2, due to improved EPB management, the first reduction in volume loss was seen (figure 2)
In Zone 3, the TBM passed under an area of heavy surface traffic where settlement was to be reduced to a minimum and the utmost care had to be taken. Very low soil volume losses in the range of 0.3% to 0.4% were recorded (figure 3).
These results were achieved by adjusting the following various working parameters.
Working pressure
The contractor, supported by the engineer, defined an EPB pressure working range. A lower limit was defined by face stability (plus a suitable safety factor) and the higher limit defined to avoid surface heave, taking into account the soil friction angle and the water table level.
As shown in figure 4, in Zone 1 the operator kept the pressure at 55-60% of the range, then increased it to 85% in Zone 2, to reach the maximum (between 95% and 100%) in the critical Zone 3.
As previously highlighted, it was also mandatory to keep the pressure in the chamber as even as possible, even during ring erection when air pressure dissipated due to foam injection (with pressure drops of 0.3 bar or more). When the TBM is at a standstill, even these small differences could have a relaxing effect on the granular soils, which would inevitably lead to surface settlements. Therefore during any planned or sudden stoppage (ring erection or belt stop) the bentonite slurry injection system was used to keep the pressure in the chamber at the designed level.
Density control of mixed soil in the chamber was also of the utmost importance, in order to keep a similar pressure gradient to the soil; moreover, it was also a clear index of the soil mix quality and the consistency of foam injection.
In the considered zones, an average density of 15.5kN/m3 was successfully maintained. In Zone 3 a slightly lower density was utilised, in order to reduce the required torque, to advance with crown EPB pressures in the range of 2.1 bar (figure 5).
The theoretical tail shield mortar volume required was 7.2m3 per ring, this quantity was evenly injected through six lines installed in the tail shield.
A few basic rules were followed to achieve effective injection. One of the most important being that injection should be pressure driven, not volume driven, as the annulus is filled simultaneously with the advance. Good practice states that the pressure of injection should be 0.5 bars, on top of the EPB pressure, this being the upper working pressure limit of the tail shield brushes.
It was also extremely important to inject from all six of the tail shield lines, to achieve even volume and pressure distribution around the ring. On this basis, advance of the TBM was halted when less than four of the grout lines were available for injection. No advance was permitted if either of the two crown lines (A1 and A6) were plugged (figure 6). Following the same practice, the number of working lines was kept as high as possible (figure 7).
Injection pressure also had its effect. This can be seen in figure 8, which compares the crown EPB pressure (P1) with the pressure reading in the top grouting lines (A1 and A6).
Slurry injection through the shield, to compensate for volume losses due to its conical shape, was also successfully implemented in Zone 3 helping to further minimise the Vl. The theoretical injection volume per ring was calculated as the difference between the excavation diameter and the tail shield diameter. Injection was carried out with seven volumetric pumps evenly distributing the 12% bentonite slurry onto the top of the shield.
The bentonite injection was not perfect, as there was a learning curve to trim the system and achieve optimum operation. Notwithstanding, this was a very important factor in reaching the reduced volume losses in Zone 3 (figure 9).
Conclusions
With the implementation of these measures, it was possible to reduce the volume loss three-fold from 1.5% to 0.5%, or even less.
This achievement was possible only due to the joint efforts of all the parties involved – client, engineer and contracting JV – by continuously improving the excavation and injection methods to reduce the recorded surface volume losses.
This valuable experience will now form the basis for the design of the next section of the line.
The 09.4m Herrenknecht EPBM at FIRA station Hany pumps installed in the back-up gantry Fig 8 & Fig 9 Fig 1 – Fig 5
