Impact of underground construction and TBM ...

Seminar on Kolkata East-West Metro.

Impact of underground construction and TBM tunnelling on surface Structures in Kolkata East-West Metro. Dr. Alvaro Casasus Associate Director, Geotechnical, AECOM, Gurgaon, India. Harvinder Rana Design Engineer, Geotechnical, AECOM, Gurgaon, India.

ABSTRACT: Kolkata East West Metro includes the contract UG-1. Design and Construction of Underground Section from Howrah Maidan Station to West end of Central Station in Kolkata. The project includes two bored tunnels mined with two Earth Pressure Balanced TBMs and with a drive length of 3,100 meters approximately. The ground is made of one upper layer of soft clay and another deeper one of stiff clay presenting also some silty sandy pockets. Through that ground the TBMs mined at depths varying from 10 to 30 meters depth and underneath various structures, including flyovers with deep foundations, rail tracks, dilapidated buildings and heritage structures. This paper summarises and reviews the methodology followed to evaluate the impact caused by the tunnel excavation on those structure, the problem faced during mining and preventive measures taken on site to minimize the structural impact.

1 INTRODUCTION

required in order take the proper measures before, during and after mining the tunnel.

Excavating tunnels in clay is a challenging work, especially when the tunnels are crossing an old city like Kolkata, having many ancient buildings, old structures and monuments. Also it is a very congested city, with traffic and people on the streets which makes difficult the access with machinery for any intervention from the ground surface. As far as the ground is concerned, the clay is a soft material which has a visco-elastic-plastic behaviour, due to the consolidation phenomenon, and presenting, most of the time, a low Young’s modulus which makes it prone to high strains under stresses. Mining tunnels through such material makes it necessary to take into account the settlements that happen, in the influence area, due to tunnel construction. These Settlements will be time dependent, and will affect the adjacent structures. That affection will depend on the stiffness of the clay and how efficiently the control of the TBM parameters is done. In any case an impact assessment of the structures has to be done to know in advance how much the ground settlement will be damaging the structures. Thus classification this damage with some engineering approaches

2 GEOLOGICAL AND GEOTECHNICAL PROPERTIES OF KOLKATA EAST WEST METRO Kolkata is located over the "Bengal basin", a pericratonic tertiary basin. The quaternary sediments in Kolkata consist of clay, silt, and several grades of sand and gravel. These sediments are sandwiched between two clay beds: the lower one at a depth of 250–650 meters; the upper one 10– 40 meters of thickness. The East West Metro alignment is located in the upper clay layer, and the ground is divided into four layers, the first one is called “made ground” and presents a thickness of 2 to 5 meters. This is classified as Unit 1 and is it formed by cohesion less material with 12Mpa of E modulus and 28 degrees of friction angle. The second layer is soft clay, classified as Unit 2, it is an organic silty clay/ clayey silt material, with a thickness of around 15 meters and a long term E modulus of 8 Mpa. The third layer it is formed by a stiff clay, it is a silty clay/clayey silt material, this layer goes 1


up to 40 meters deep and it is classified as Unit 3A. The drained E modulus of this layer is equal to 55Mpa at average. In the alignment there are also some pockets of silty sandy material, Unit 3B, with same E modulus as Unit 3A but higher permeability. In the figure 2 it can be seen a general scheme of the ground.

diameter, tunnel excavated diameters, is equal to 6,380 millimetres (mm), while the tail of the shield has a diameter of 6,330 mm, hence from the cutter head to the tail shield there are 25 millimetres of gap in which the ground can close freely. The reason of having a tapered shield instead of a straight one is to reduce the ground friction and the possibility of getting trapped in case the ground closes around it. Finally the segmental lining ring is placed right at the end of the shield, that ring has an external diameter of 6,100 mm, so from the shield tail to the outer face of the segmental lining there is another gap of 115 mm. This gap is filled with what it is called, liquid A and liquid B. Liquid A is a cement grout and liquid B sodium silicate to accelerate the gel time of the grout, to give confinement to the segment and stop deformations as soon as possible. Hence can be assumed that the last 115 mm of gap will not create any volume loss, since it is filled with the cementitious grout, however it may also produce some ground loss because of the grout shrinkage and improper filling of the gap, this loss is termed as tail loss. There is another volume loss to take into account, the face excavation volume loss, once the tunnel is excavated the ground on the excavated face will try to extrude inside the tunnel tube. This phenomena can be controlled as long as the proper face pressure is used.

Figure: 2 Ground profile Kolkata East West Metro

3 CONCEPT OF VOLUME LOSS When a continuous ground is excavated, its in situ stresses redistributes along the periphery of the excavated area, experiencing a reduction on radial stresses and an increment on tangential ones, this stresses redistribution will create radial strains which will cause deformations which we know as convergences. The ground will close inwards to the opening and this closure is what we define as volume loss. If this phenomenon is allowed without any support, the ground will keep closing until it reaches a new stresses in equilibrium state. However, if the ground strength parameters are not able to resist those stresses then the excavated opening will collapse. In tunnelling what is done is; we place a support/lining to resist the percentage of stresses, which the ground cannot resist. But there is a gap of time and space in between the moment when the tunnel is excavated and the lining/support is placed. During this mentioned gap, the ground is free to deform and volume loss will take place in this period of time. In the case of tunnels mined with a TBM machine, Earth Pressure Balanced (EPB), like in the Kolkata East West Metro, the spatial gap in between the ground excavation and installation of the support, is equal to the length of the shield. In our particular case it is a Herrenknecht EPB with a shield length of 7,594 millimetres. On the other hand the cutter head

Figure 1: Volume loss scheme along the TBM

So we can conclude here is; the only volume loss which cannot be controlled in an EPB machine is the shield loss. The convergences that take place from the cutter head up to the shield tail in the case of the EPB used in the Kolkata East West Metro UG-1 can be up to a maximum of 25 mm. This figure represents a volume loss of 1.56%. During the impact assessment on the structures a volume loss of 1.5% was considered systematically. The volume loss in the excavated face can contribute around 0.1% to 0.2% more, as 2


mentioned above it will depend very much on the operational face pressure the EPB is using. Having a gap of 25 mm in which the ground can freely close around the shield does not mean it will do it totally, that will depend on the ground parameters and stresses. Dr. Kirsch (1898) presented the elastic solution for the stresses around a circular hole made in a loaded infinite plate. Later Brady – Brown, based in Kirsch solutions, developed formulations for calculating displacements around the circular opening.

(1)

=

[(1 + )

(

)(

mm, that represents a volume loss of 0.83%, and the ground will not close around the shield, being the final gap to fill with grout equal to 115 + 25 19 = 121 millimeters. On the other hand what we have analysed so far are long term or drained E modulus parameters for the different layers, here we are going to discuss what happens with the short term or undrained parameters. As it is well known that the deformations on clays have a time dependent behaviour and that happens because of the consolidation phenomenon. According with Karl von Terzaghi's Principle, in clays when there is an increment of stresses at the very beginning this increment will be producing an increment in the pore pressure only, remaining the effective stresses constant. However little by little the stresses from the water are dissipated into the ground particles, increasing the effective stresses and the deformations. That phenomenon is known as consolidation and the time it takes to happened depends on the ground permeability and the level of stresses. The conclusion is that the deformation of the clays is also linked with the time the clays remain unsupported. If we consider undrained or short term parameters for the Kolkata clays then we will have a reduction on 25% of the deformations. Since according with the GIR the undrained E modulus is 25% higher than the drained one. Considering this maximum deformation of 42 and 16 mm in Units 2 and 3A can be calculated respectively.

4 ) cos 2 ]

This formulation can be used for computations of the maximum radial displacement as per ground parameters and stresses and the maximum volume loss, by integration of those displacements along the circular opening. If the displacements are higher than 25 mm then the ground will close around the shield, acting this one as temporally supported until the segmental ring is placed and the gap is filled with cement grout. If the gap is filled properly more displacements of the ground should not happen, j ust the one equivalent to the grout shrinkage, which is not more than a few mi llimetres. Considering 12 meters depth in Unit 2 and the long term E modulus, applying equation (1) we obtain a maximum radial deformation of the tunnel equal to 53 mm. Here we have to call the attention on the elastic analysis done; we are not taking into account the plastic deformation of the clay. These 53 mm deformation implies that the ground will close around the shield and the shield will act as temporally support until the ground reaches the shield tail where the segmental ring will be installed and the 115 mm gap will be filled with the cement grout. As already it has been indicated, if the void is properly filled, not more as 25 mm ground inwards deformation should happens and the volume loss would be around 1.5%. Now if we consider the layer Unit 3A, as already discussed, the long term E modulus is equal to 55 Mpa, doing the same exercise, for a depth of 25 meters we can expect a maximum elastic deformation of the ground equal to 19

4 PREDICTION OF GROUND SETTLEMENTS ON THE SURFACE The volume loss created during boring of the tunnel will end up creating ground settlements on the surface. Those settlements will depend on the quantity of volume loss and type of ground. For that prediction, different software based on Finite Element or Finite Differences approach or an accepted empirical method based on the Gaussian curve settlement distribution proposed by Peck (1969) can be used.

= 3

(2)


each tunnel independently and adding the settlements, represented by each Gaussian curve, in the overlapping zone. Following the previous calculations the Gaussian curves for two tunnels mined in Unit 2, at 12 meters depth is presented in Figure 4, and two tunnels in Unit 3A at 25 meters depth is presented in Figure 5.

The representation of equation (2) can be seen in Figure 3.

Figure: 3 Ground settlements profile

Integrating the equation (2) we obtain the total volume of settlements on the surface, which is equal to the tunnel volume loss.

=

(%)

4

=

Figure: 4 Gaussian curves for two tunnels mined in Unit 2 (12 m Deep)

Finally we obtain the expression of the maximum settlement on the surface which it is linked to the ground volume loss.

(3)

=

0.313 (%)

The parameter represents the distance of the Gaussian curve inflexion point to the tunnel axis, and it is equal to: (4)

is a parameter depending on the type of ground. For the Kolkata East West Metro, contract UG-1 impact assessment, that value was taken as 0.5, being as per the literature the most suitable value for clays. That value was checked afterwards with the actual ground settlements realizing that was it was fitting quite well the settlements points in the theoretical Gaussian curve sometimes. However in some stretches was found that smaller values were better fit. To take into account the settlements produced by the two tunnels the methodology followed was to consider the Gaussian curve of

Figure: 5 Gaussian curves for two tunnels mined in Unit 3A (25 m Deep)

We can draw the conclusion that in Kolkata East West Metro, contract UG-1, the ground settlements caused because of twin tunnels excavation, will be expected in the range of 15 to 30 millimetres.

4


Equations (5) and (6) allow calculating the shear and bending stresses on the building once the ground settlements are predicted. The methodology is finalized categorising the building damage depending on the maximum composed tensile strain imposed to the building. In the Figure 7 it can be seen that classification.

5 IMPACT ASSESSMENT OF GROUND SETTLEMENTS ON THE STRUCTURES In Kolkata Metro tunnelling impact assessment was done for all the structures falling on the tunnels influence area. In that area both shallow structures foundations, heritage and old dilapidated buildings, as well as deep structures foundations, flyovers and other buildings were assessed. For that assessment both analytical and numerical approaches were used.

5.1 Assessment of Structures on Shallow Foundations The analytical approach followed was based on J. B. Burland et al. (2002) Ref.1, and Marco D. Boscardin et al. (1989) Ref. 2. The assessment methodology is based on representing the building, affected by the settlements, as a thick Timoshenko Beam; shear deformation is not neglected. The beam will be simple supported in both ends and a punctual load will be applied to the centre. This approach gives reasonable estimation of strains on the building based on the displacement of ground and it has been proven by Boscardin and Cording with results from field tests.

Figure: 7 Damage chart based on maximum strain.

According with the methodology from categories Negligible to slight no further study of the impact is required. However from Moderate to Very Severe a detailed analysis of the tunnel structure interaction is required and measures to reduce the impact has to be put in place. It should be noted that this approach considers the tunnel passing right below the building which is not always the case. While assessing the impact of the tunnels on the influence area located structures, in Kolkata Metro, it was realized that some of them were in a different position, with the tunnel crossing below on the structures edges. Under those circumstances applying Burland’s equations (5) and (6) in an accurate way is complicated, so another formulation, based in the same approach was develop by General Consultant team (GC), in other to check the impact on those structures. The methodology is based on considering the buildings as a thick beam fixed in both ends and applying a vertical displacement equal to the ground settlement in one of them.

Figure: 6 Schematic Representation of a building for impact assessment.

The assessment is based on calculating the shear and tensile stresses imposed on the building because of the beam deflection that is, the ground settlements.

=

(5)

(6)

=

1+ 12

+

18 2

3

Figure: 8 Representation of building as fixed beam.

5


negative friction will happens, the ground will slide along the pile imposing on this an extra axial force, equal to the limit ground – pile adhesion force; , and reducing the length of pile resisting the foundation forces, that will end up in increasing the pile settlements. Determining the length of pile in which the negative friction will take place is the right way of determining the settlements on piles. In the figure below it can be seen the phenomenon in which the pile has been considered as a rigid body; elastic deformations on it had been neglected. The circle is showing the pile length in which negative friction will take place.

The formulation we arrive for shear and bending strain can be found below:

(7)

(8)

8 + 5 +

5.2 Assessment of Structures on Deep Foundations Tunnelling in the vicinity of piles will impact those in two different ways, one will be structurally impact, increasing bending moments and axial forces, and the other will be affecting the bearing capacity of the pile through the ground friction reduction. The impact will depend on the position of the pile with respect the tunnel centre, when the tunnel passes beside the pile, it will deform the ground creating bending moments on the piles because of the imposed displacements. Also a reduction in friction because of horizontal stresses relaxation will take place, reducing the pile capacity, H. G. Poulus Ref. 3. That lost in capacity may end up in settlements on the plies.

Figure: 10 Ground settlements profile versus pile settlement. Negative friction.

6 IMPACT ASSESSMENT ON KOLKATA EAST WEST METRO UG-1 In this section the impact assessment done in the most important structures along the tunnel alignment in Kolkata East West Metro UG-1 is detailed. Colvin Court Building The building is basically load bearing brick masonry structure with steel beam supporting the floor slab. This G+3 storied building, with shallow foundations, is mainly for Eastern Railway’s senior officer’s Bungalows.

Figure: 9 Lateral impact of a tunnel in piles.

However assessing the settlements that piles founded structure will suffer because of the excavation of a tunnel is not an easy task. The main difficulty it’s in the understanding and quantification of the pile – ground interaction. Using numerical models will overestimate those settlements unless the code can simulate the ground sliding along the pile. Here comes the concept of negative friction, the movements of the pile and the ground will be different, whenever the vertical displacement of the ground is higher than the one of the pile, then

Picture: 1 Colvin Court front view

6


Figure: 11 Position of tunnels with respect Bankim Setu Bridge piles.

Figure: 10 Plant view tunnels and Colvin Court

The tunnels crossed below the building at a depth of 14.5 meters, inside the Unit 2 layer. For the impact assessment 1.5% of volume loss was considered and a maximum of 30 millimetres settlements was expected with a negligible damage in the building. However the TBMs were in the “learning curve” at that moment, during that initial situation the TBM parameters are not yet adjust in the optimum way. The final maximum settlement recorded was 70 millimetres, around 3.5% of volume loss. The damages on the building were under slight category in the end, some cracks, less than 5 millimetres thick, were reported. Bankim Setu Bridge This is a RCC bridge which crosses the Buckland road over the train tracks, and it is founded on 25 meters long piles.

To assess the impact of the tunnels on the structure a FEM model was done using Plaxis 2D. The results on the model predicted a single pile maximum settlement of 15 millimetres, while the actual settlement after the tunnels crossed the area was 5 millimetres maximum. That shows the, already discussed, limitations of the FEM software to simulate properly the interaction between pile and ground, not being possible to simulate the detaching of the ground from the pile structure in the negative friction length. At this stage TBMs were running smoothly, after the learning period, and settlements on the ground were showing volume losses less than 1.5%. No damage was reported on the referred bridge. However as precaution the bridge was closed to the traffic during the period of time the TBMs were mining beneath. DRM Building This is a G+2 building, structure, which is formed by load bearing external brick masonry walls with steel column inside the building supporting steel beams and floor slabs, the foundation consists on a continuous shallow foundation. The building is mainly occupied by GRPF police station at ground floor level and DRM office at first and second floor level.

Picture: 2 Bankim Setu Bridge

The tunnels cross below the bridge at 24 meters below the ground level and the nearest pile to the tunnels is at 3.83 meters of horizontal distance. The tunnels are in the Unit 3A, the stiff clay with SPT around 30, E modulus around 50 Mpa and 250Kpa of undrained shear strength.

Figure: 12 Plan view of DRM building respect tunnels.

7


The tunnels depth along the building varies from 27 to 30 meters and the layer in which the tunnels mined is Unit 3A, stiff clay with SPT values up to 50.

simply supported made up of composite structure, steel and concrete.

Figure: 14 Plan view of Brabourne Road Flyover respect tunnels.

Picture: 3 DRM Building view

GC did the impact assessment of this structure, predicting total ground settlements of around 30 millimetres and damage on the building in the very slight category.

The effect of tunnelling on the pile foundations has been simulated with 2D finite element modelling approaches. The analysis was carried out for different sections under Brabourne flyover to assess the settlements on the flyover piles. The ground parameters at this point belong to very stiff clay with SPT values up to 70.

Figure: 13 Settlements prediction below DRM building.

Even though predicted settlements were a bit high, the position of the tunnels were favourable, the differential settlements suffered by the building were around 3.5 millimetres. Because of operational constrains this building could not be evacuated during the mining process, however preventive measures were implemented like horizontal nets to prevent any spalls from falling down. The TBMs crossed the building registering settlements as predicted and no major damages on the building were reported. Brabourne Road Flyover This flyover is located along the tunnel alignment after Hoogly River. The tunnels pass beneath 6 piers of the road flyover. The ground layer crossed is stiff clay, Unit 3A, in this section. The crown of the tunnel at this location varies from 30 to 32 meters deep. Clear spacing between the pile bottom and tunnel crown varies from 7.5 meters to 9 meters. The superstructure of Brabourne Flyover Bridge is

Figure: 15 Brabourne Road Flyover cross section

From the analysis it was observed that the maximum settlement in the pile is equivalent to the maximum settlements in the ground which was predicted to be 20 mm.

8


Figure: 16 FEM analyses for piles settlement

However after the TBMs crossed below the flyover, a maximum of 10 mm settlement was observed during real time ground monitoring. The software predicted higher values of settlements in the piles as it does not allow the piles to detach from the soil. This results in the movement of piles together with the mesh resulting in higher settlements. The TBMs crossed the bridge without any damages reported. However as precaution the bridge was cut to the traffic during mining. Dilapidated buildings along Raja Woodmount St. There are some buildings nearby Brabourne Road Flyover which were a concern during mining beneath them because of the poor condition of them. The structure of those buildings was in a quite bad state being the buildings a risk itself before the tunnels crossed below them.

Picture: 4 Dilapidated Woodmount St. view

buildings

along

Raja

Writers Building and St. Andrews Church At the time of writing this paper there is still a challenge pending in the project, and this is crossing the heritage buildings; Writers and St. Andrews Church. Writers building is an edifice with a great political significant and memories of Indian Independence Movement. Both buildings are founded with shallow foundations 1.5 to 2.0 meters deep. The twin tunnels will pass beneath the buildings edges at a depth of 17 meters from the ground surface before entering in New Mahakaran Station.

Figure: 17 Plan view of Dilapidated buildings along Raja Woodmount St.

Also the ground layer crossed was stiff clay, Unit 3A, in this section. The crown of the tunnel at this location varies from 30 to 32 meters deep. The previous assessments of these buildings showed maximum settlements of 20 millimetres, with 1.5% volume loss, and an impact category of negligible. However because of the poor condition of the building evacuating them during the mining process was considered as a protective measure. The actual settlements recorded after the TBMs crossed was up to 7 millimetres, and no damage on the buildings, beyond the ones they already had before the TBMs, was reported. That means a volume loss around 0.7%, and it can be back calculated an E modulus of around 90 Mpa. that gives an idea about the stiffness of the clay in that stretch.

Figure: 18 Plan view of Writers and St. Andrews Church with respect to the tunnels.

At this location the TBMs will be mining partially in Unit 3A, 3B and Unit 2 on the tunnel crown. The SPT values are ranging from 10 in the Unit 2 up to 30 in Unit 3A. The contractor has prepared a FEM using Plaxis 2D to assess the expected settlements in Writers Building, from this model they have obtained a maximum settlement of around 22 millimetres on the ground surface. With those settlements they have performed an impact assessment obtaining a maximum tensile stress of 0.114% falling the building under the slight category of damage. That means that non9


injecting a low pressure grout, 1 to 2 bars with a cement water ratio of 0.5 in weight. This measure was applied in all the critical buildings showing good results.

structural cracks on the building are expected, being any possible damage easily repair afterwards.

Figure: 20 Peripheral grouting scheme.

Figure: 19 Writers Building FEM settlements assessment.

7 PRE-EMPTIVE AND MIITGATION MEASSURES During mining below all the critical buildings and structures some pre-emptive and mitigation measures were put in place to minimize the settlements and the consequences of those in the structures. Summarized of these measures is presented below: 1. Reducing time; it was quickly learnt the time dependant behaviour of the clay, finding that reducing the gap of time between excavating a ring and placing the lining the settlements were also reduced. That happened especially in the unit 2. 2. According with previous point, stoppages of TBM should be avoided in the vicinity of any sensitive structure. 3. TBM parameters has to be control in a very close manner, face pressure, volume of material excavated and volume and pressure of grout injected in the annular gap are directly related with the settlements on the ground surface. 4. Propping in buildings, this measure consist on place some propping in arches, windows, big structural spans, near pillars,… the main objective of these is to make the building more rigid as structure not following the ground settlements and reducing the cracking. This measure is complemented with the next one. 5. Peripheral surface grouting along the buildings, this measure will aim to fill any gap which can exist or create during/after mining in between the ground and the buildings foundation. It consists on

6. Buildings evacuation and bridges traffic cut; whenever was possible all the critical buildings crossed by the TBMs were evacuating while the machines were mining and all the bridges were cut to the traffic. 7. Continuous monitoring of ground and structures movements during mining; that was done during the whole mining process and in almost real time for the most critical structures.

8 REFERENCES Ref 1. Burland J.B., Standing J.R. and Jardine F.M. (eds): Building response to tunneling. Case studies from the Jubilee Line Extension, London. Ref 2. Boscardin, M.D. and Cording E.G. (1989): Building Response to Excavation Induced Settlement. Ref 3. Poulos, H.G. and Deng, W. (2004). An Investigation on Tunnelling-Induced Reduction of Pile Geotechnical Capacity. Ref 4. Kirsch G. Die theorie der elastizitat und die bedurfnisse der festigkeitslehre. Veit Ver Deut Ing 1898; 42: 797-807.

10