ABSTRACT
Construction of urban tunnels in soft soil or weak rocks requires meticulous considerations in terms of geotechnical site investigations, construction methods, types of tunnel boring machines, tunnel support systems, instrumentation and monitoring of surface subsidence and the subsequent impact on nearby buildings and services. Among the considerations, the most important aspect is the control of surface subsidence to minimize any disturbance to nearby buildings and services. The study of surface subsidence is limited to mainly empirical solutions based on field studies, and very few analytical studies have been carried out. The available analytical solutions are not sufficient to include complex ground conditions and hence a comprehensive analytical solution coupled with numerical modeling is necessary to model the effect of surface subsidence due to soft ground tunneling. This paper discusses a variety of parameters and considerations that are required to formulate a theoretical solution to predict surface, sub surface and lateral settlement due to tunneling induced ground movement. Moreover, the importance of considering the tunnel support system as a parameter in determining the settlements is often neglected or excluded and thus is highlighted in the paper. A number of empirical solutions to predict the settlements are also compared to differentiate their effectiveness in various soil conditions.
Keywords: soft ground tunneling; soft soil; weak rocks; subsidence; tunnel support system.
INTRODUCTION
As infrastructure, buildings and services stretch through the densely populated and scarcely limited land space, ensuring minimum disturbance to the routine daily lives of the city is of utmost importance in tunnel design. In order to fully understand the extent of disturbance due to tunneling in such tight conditions, a comprehensive knowledge of the deformation caused by tunneling will be useful. Over the years, tunnel engineers and researchers have realized that curbing the loss of ground is futile because of in-situ conditions and ground ambiguities. Hence the only solution is to control the effects of deformation rather than to totally stop it. Through various field studies and experiences, empirical methods have been developed over the years to predict the settlement caused by tunneling in soft ground by Peck (1969), Attewell et al. (1986), New and O’Reilly (1991) and semi empirical methods were also developed by Lo et al. (1984). Some important considerations for the prediction of settlement are presented in their solutions but may not suffice in obtaining a total solution.
Continuous research and advancement in technology towards tunneling works will inevitably lead to safer and both economically and environmentally efficient construction process. Besides obtaining field data to formulate empirical relationships of ground deformation, a major difficulty is the inconsistency of soil condition and applicability of the empirical formulas to different type of soil. Hence, an analytical solution is required to cater to all soil type and varying site conditions. In order to formulate an analytical solution to represent ground surface and sub surface deformation, several factors and parameters relating to soil type and characteristics, tunnel geometry and specifications, type of tunnel boring machines (or TBMs) and advance rate, and type of tunnel support are primarily required. This paper will discuss a variety of parameters and factors that will affect ground surface settlement as a result of tunneling induced ground movement.
GROUND SUBSIDENCE
Surface settlement
Limiting the surface settlement produced by tunneling in shallow and soft ground is the major challenge of any tunnel engineer. Surface subsidence can be caused by several factors and some of them have been discussed earlier such as ground loss at the tunnel face, behind the tail of the shield and through the tunnel support or linings. Other factors include the consolidation of the soil due to reduction of ground water level. Besides surface settlement, tunneling also produce lateral deformation of the ground and longitudinal movement of the ground ahead of the tunnel face. Little studies have been devoted to the theoretical aspects of deformations caused by tunneling. Most methods derived for estimating surface or subsurface settlements are empirical in nature and based on field observations of mainly UK tunnels. The most common method is by estimating the value of i, a parameter used to define the distance from the tunnel centre line to the point of inflexion of the settlement trough of a normal probability curve as shown in Figure 1. The distribution of the settlements or settlement trough at the level of the foundations approximated a normal probability distribution function (Equation 1) described by Peck (1969).
![]() |
(1) |
where w = surface settlement at a transverse distance (y) from the tunnel centre line
wmax = maximum settlement at x = 0
i = location of maximum settlement gradient or point of inflexion.
New and O’Reilly (1991), Mair et al. (1996) and Attewell et al. (1986) have described the application of the Gaussian distribution to the prediction of vertical and horizontal movements due to single and multiple tunnels. Bracegirdle and Mair (1996) have also used the normal distribution function to evaluate the potential damage to cast iron services or pipes induced by tunneling. Similarly, Yoshida and Kusakabe (1994) described the behavior of ground and adjacent underground piping during shield tunneling. In addition, Durand et al. (1994) used the normal distribution function in their design approach for the Toulon underground motorway crossing to predict surface settlements and their effects on nearby buildings. Values of i have been calculated for tunnels based on reasonably reliable settlement data. The calculated results have been assembled and illustrated in a dimensionless plot of i/R against z/2R for various tunnels in different materials by Peck (1969). Using the relationship obtained from field observations, Peck (1969) formed the following relationship to estimate the value of i,
![]() |
(2) |
where n = 0.8 to 1.0 and zo is the tunnel depth above the tunnel crown
Figure 1. Surface settlement profile.
Table 1. Different empirical solutions of settlement trough, i.
![]() | Reference | i | Basis for empirical solution |
![]() | Peck (1969) | i/R = (zo/2R)n (n=0.8 - 1.0) | Field observations |
Attewell and Farmer (1974) | i/R = (zo/2R)n (a=1, n=1) | Field observations of UK tunnels |
Clough and Schmidt (1981) | i/R = (zo/2R)n (a=1, n=0.8) | Field observations of UK tunnels |
O’Reilly and New (1982) | i = 0.43zo + 1.1m (cohesive soil (3 £ zo £ 34m)) i = 0.43zo + 1.1m (cohesive soil (3 £ zo = 34m)) i = 0.28zo - 0.1m (granular soil (6 £ z0 £ 10m)) |
Field observations of UK tunnels |
Atkinson and Potts (1977) | i = 0.25(zo + R) (loose sand) i = 0.25(zo + R) (loose sand) i = 0.25(1.5zo + 0.5R) (dense and OC clay) |
Field observations and model tests |
Leach (1985) | i = (0.57 + 0.45zo) ±1.01m | For sites where consolidation effects are insignificant |
Mair et al. (1983) | i = 0.5zo | Field observations and centrifuge tests |
![]() |
A comparison of the various empirical methods discussed above was made on the assumption of a hypothetical four meter diameter tunnel located at a depth of thirty meters which experience a ground loss volume of one percent. The results are shown graphically in Figure 2.
Figure 2. Comparison of surface settlement troughs.
From the comparison of various empirical solutions for surface settlement trough, the maximum settlement ranges from 3-5 mm whereas the trough width (i) varies between 10 and 15 m. This shows that there are significant discrepancies between empirical solutions to predict surface settlement trough because of different interpretation and database collection proposed by different authors.
Longitudinal settlement
Very limited analytical studies have been conducted to understand the longitudinal behavior of the ground along the tunnel axis because field studies using extensometers or probes experience difficulty in installation thus few intensive monitoring of this behavior had been carried out. Attewell and Woodman (1982) assumed that the longitudinal profile of tunneling induced settlement could be modeled as a cumulative probability curve and drew a conclusion that the settlement directly above the tunnel face is half of Smax, which is the maximum transverse surface settlement.
However, Attewell et al. (1986) conducted lab extrusion tests using undisturbed clay soil forced through a hole whilst applying a constant axial pressure to analyze the intrusion at the tunnel face. The extrusion displacement profile developed is shown in Figure 3. Romo and Diaz (1981) used finite element method to find the stress and deformation of different degrees of face yielding and established a relationship between safety factor, stability ratio and surface settlement at different depths that are attributable to face yielding. The settlement along a tunnel alignment due to face yielding as the tunnel progressed between two points can be calculated by Equation 3.
![]() |
(3) |
where wx = the settlement at a distance x from initial point of reference,
Z = H + D/2,
H = tunnel depth above the crown,
D = tunnel diameter,
sh = initial horizontal stress at tunnel axis,
sf = mean compressive soil strength from ground surface to depth of tunnel invert,
p = fluid pressure at excavation face,
ef = mean axial strain at failure of soil samples from ground surface to depth of tunnel invert and
F1 = a function related to x/(Z + D).
Figure 3. Longitudinal profile of surface settlement (Attewell et al., 1986).
Lateral settlement
Studies involving the lateral movement of the ground due to tunneling were comparatively more extensive than studies conducted on the longitudinal behavior. Norgrove et al. (1979) derived an empirical equation that relate the sub surface settlement to the lateral deformations as shown below
![]() |
(4) |
where wy is the lateral deformation, w is the surface settlement at a distance y from the tunnel axis, zo is the tunnel depth above the tunnel crown and y is the horizontal distance from the tunnel axis. However, these empirical equations are subjected to certain limitations and assumptions in terms of ground applicability, excavation methods and inadequate site information, thus making them inaccurate to a certain extent. O’Reilly and New (1982) assumed that the resultant vectors of ground movements are directed to the tunnel axis and proposed an empirical similar equation to Norgrove et al. (1979) with the vertical and horizontal components of the ground movements as Sv and Sh, and the horizontal surface settlement can be calculated as
![]() |
(5) |
where Sv is the settlement at a distance y from the tunnel axis, y is the horizontal distance from the tunnel axis and zo is the tunnel depth above the tunnel crown.
Theoretical solutions
The normal probability curve proposed by Peck (1969) to model the observed profile had no theoretical basis and solely based on the fact that the curves resembled each other in shape. The normal probability curve can be effectively used to predict the actual surface settlement profile but it does underestimate the actual deformations when applied to granular soils (New and O’Reilly, 1991) and to overconsolidated clays (Eisentien et al., 1981). In addition, the normal probability profile does not provide intuitive solutions for the sub surface and lateral soil movements due to tunneling.
A small number of authors presented analytical predictions of tunneling induced ground movement in clay. The most significant is by Sagaseta (1987) who proposed closed form solutions for obtaining the strain field in incompressible soil by combining fluid flow with elastic solutions for half space. Sagaseta (1987) solved for a singularity at a point of an elastic half plane and by adding the image solution for the singularity at a point located symmetrically above the soil surface, the normal or shear stresses are made to neutralize. A negative mirror image of the point with respect to the top surface will produce opposite normal stresses and the same shear stresses similar to the actual point. Conversely, a negative image will produce the same normal stresses and opposite shear stresses. However, these elastic solutions might not be effectively applicable for soft ground since elastic solutions are more applicable for rock. Verruijt and Booker (1996) modified the elastic solutions by Sagaseta (1987) and applied various values of Poisson’s ratio and included the effect of ovalization of the tunnel opening. However, the solutions yielded slightly wider profiles of lateral deformations induced by tunneling with tunnel boring machines. Hence, Loganathan and Poulos (1998) redefined the ground loss parameter with respect to gap parameter and used it in the solutions provided by Verruijt and Booker (1996). Their results yielded more accurate profiles of surface, sub surface and lateral deformations as illustrated in five case studies provided by Loganathan and Poulos (1998). The gap parameter can be estimated based on a theoretical method developed by Lee et al. (1992) as given below
![]() |
(6) |
![]() |
The generalized modified analytical equations based on Verruijt and Booker (1996) for the estimation of the surface settlement (Uz=0), sub surface settlement (Uz) and the lateral deformation (Ux) are:
![]() |
(7) |
![]() |
(8) |
![]() |
(9) |
The ground deformation due to the ovalization term was deliberately omitted since the analysis was concerned only with short term undrained conditions. According to Strack and Verruijt (2000), ovalization of a shallow tunnel is unlikely to exist since the stresses accommodating an ovalization of the tunnel cavity did not correspond to an ovalization of the tunnel lining. Figure 4 illustrates the boundary conditions used in the derivation of the ground loss.
The theoretical solutions provided by Sagaseta (1986), on which other authors modified to predict soft ground deformations due to tunneling, is essentially based on incompressible soils. Hence, it might not accurately predict the deformations in soft ground. Elastic solutions are more applicable for hard rock conditions.
Figure 4. Ground deformation patterns and ground lost boundary conditions (Loganathan et al, 2001)
TUNNEL FACE STABILITY
Tunnel face stability were investigated by Broms and Bennermark (1967), who had proposed using a face stability index (N) which is the ratio between the difference of the natural pressure and the pressure applied to the tunnel face, and the undrained shear strength to analyze tunnel face stability. The stability ratio is devised based on cohesive ground (clay) and it was found that the tunnel face would be stable when the index is less than six.
![]() |
(10) |
The claims made by Broms and Bennermark (1967) of N<6 for stability of tunnel face were substantiated by Mair (1979), Schofield (1980) and Davis et al. (1980) who concluded that the stability ratio depends on the depth of cover to diameter (H/D) ratio and that stability ratio (N) is equal to between five and seven for depth of cover to diameter ratio of 1.5. However, Davis et al. (1980) found that their results estimated from the limit theorems of plasticity yield critical values of N that showed a significant variation of the depth of cover to diameter ratio from that stated in Broms and Bennermark (1967). A larger range of N for tunnel face stability was given by Kimura and Mair (1981) who carried out tests in centrifuge on reconsolidated clay and verified that the stability of the tunnel face can be confirmed for values of N between five and ten depending on the depth.
However, Romo and Diaz (1981) felt that the stability ratio does not explicitly reflect the actual stability of the tunnel face and so they devised a safety factor (which is defined as the ratio between shear strength of the soil and maximum shear stress) for the stability ratio. A series of plots of contours and critical surfaces corresponding to a number of stability ratios and tunnel depth to diameter ratios were devised and the relationship between the safety factor and stability ratio was obtained. Results correlated well with the reported case histories from Peck (1969) whereby tunnel face failure occurs in the range between six and seven. The reported stability ratio at safety factor of one is approximately 6.5. Although their results showed good correlation with case histories by Peck (1969), it should be noted that they had assumed that the soil behaved as a non linear elastic material having a hyperbolic stress strain curve and that no slippage would occur at the shield-soil interface during the tunnel face excavation.
Tunnel face stability in sandy and pervious water bearing grounds was investigated by Leca and Dormieux (1990). They proposed a support pressure (sT) to be applied to the tunnel face using a three-dimensional failure mechanism as described in Leca and Dormieux (1990) involving the rigid body movement of 2 conical blocks. They had proposed a limiting face pressure (sT*) for dry cohesionless soils as shown in Equation 11.
![]() |
(11) |
where aS and ag are weighting factors dependent on friction angle and depth of cover to tunnel diameter ratio (H/D) and sS is the overburden pressure.
This approach was further investigated by Chambon and Corte (1994) who formed a relationship between the tunnel face axial deformation and the pressure applied to the tunnel face based on centrifuge tests of Fontainebleau sand as shown in Figure 5. They observed that the critical and collapse pressure (Pc) increases with the tunnel diameter. Einsenstein and Ezzeldine (1994) also carried out a study to investigate ground behavior with respect to stress reduction at the face of an excavation using a shield by means of two separate analyses involving an axisymmetrical and a three dimensional finite element analysis in an ideal elasto-plastic field. The results obtained were compared to results obtained from Chambon and Corte (1994) and it was found that the face pressure (Pf) (Figure 5) was very close to the value of Rankine’s active pressure.
Figure 5. Schematic trend of the tunnel face axial displacement (d)
with reference to the applied pressure (P) (after Chambon and Corte, 1994).
CONSIDERATIONS FOR PREDICTION OF TUNNELLING INDUCED GROUND MOVEMENT
It is difficult and maybe impossible to incorporate every single factor that influences deformation of the ground as a result of tunneling into a theoretical solution for predicting the settlement. The aim of this study is to present every factor that causes or controls ground deformation induced by tunneling in soft and weak ground. The following presents a comprehensive list of these influencing factors:
(a) Comprehensive site investigation (i.e., physical properties of soil and deformation parameters, permeability, composition of strata, groundwater table, existence of nearby services or pilings or existing tunnels, surface loadings, existence of rocks and boulders etc)
(b) Tunnel requirements (i.e., depth, diameter, purpose, linings, single or parallel tunnels, duration of construction, project cost etc)
(c) Construction methods (i.e., cut and cover, New Austrian Tunneling Method (NATM), tunnel boring machines such as slurry or earth pressure boring machines or open face types, pre-support reinforcements, concrete linings, tunnel support systems, tunnel face reinforcement and drainage)
(d) Instrumentations and monitoring (i.e., ground surface and lateral deformations, ground water intrusion, linings, cavity convergence, face stability and long term settlement)
(e) Environmental impact (i.e., pollution, ecological impact and health)
Site Investigations
The type of soil, grain size, density and moisture content will affect the amount of settlement induced by tunneling. Peck (1969) gave a comprehensive description of the soft soils ranging from cohesionless granular soil to saturated stiff hard clay and the corresponding settlement induced. The settlements above and lateral to the tunnels in plastic clays of soft to medium consistency are relatively larger than in stiffer, more brittle cohesive granular soils. Currently, most research works only consider a homogenous type of soil material but through experience, the site conditions usually do not hold true for this homogeneity. For example, in predicting surface settlement due to ground movement induced by tunneling, the settlement trough width is empirically calculated using solutions derived mostly from case studies of London Clay (Attewell et al., 1977; and Clough and Schmidt, 1981) and in addition, there are several different solutions for the determination of settlement trough width in London clay. Such ambiguity often leads tunnel designers to use the solution that has a good track record rather than one that really symbolizes the actual ground conditions. In the past, drill and blast methods were used when hard rocks were encountered. Nowadays, TBMs are designed specifically to work efficiently in a particular type of ground. However, in practice, the most common problem faced was encountering a mixture of soil types which would diminish the efficiency of the designed TBMs. Usually such problem occur when a TBM specifically designed to bore through soft soil suddenly encounters hard rocks or boulders. However, this could be minimized if every tunnel is bored with a road header with extremely strong drill bits. Consequently, extra cost has to be incurred and cost efficiency is compromised. These problems will surface if a thorough site investigation was not performed. Analytical studies with respect to mixed face tunneling, are also limited due to the complexity of the problem and hence most of the time, decisions are made on site when the actual problems occur. The most common solution will be to blast the rocks and drill through the softer part of the soil so that the shield can progress but shielding the soft soil whilst blasting the rocks can pose major difficulties. As such, a two phase cutting wheel, which can disintegrate the boulders and remove them from the spoils, was developed to counter the problem of mixed face tunneling where hard rocks and soft soil are encountered (Eisenstein, 1995). Rock disc cutters are incorporated ahead of the soil pick scrapers to disintegrate the rocks, which are then transported concurrently with the soil mix.
Groundwater considerations in tunnel design is an important parameter as groundwater seeps into the tunnel through the tunnel face or cavity, or water pressure on the tunnel arches might incur the problem of cavity collapse. Soil materials are washed into the tunnel and water can also cause flooding of the tunnel as a result of water inflows. This is especially true when fine grained soil clusters are encountered in clay or rock formations. In such cases, techniques to lower groundwater level must be implemented using pumps or drainage pipes. However, lowering groundwater level can cause damage to surrounding building foundations and hence extra measures should be taken prior to that. Locations of buildings, foundation piles and services have to be identified thoroughly before excavations begin. Extra support system might have to be used if tunneling is executed under a building to control any deformation. Surface settlement will also caused inconvenience to roadways or services due to deformation induced by tunneling works as seen in reports by Attewell and Farmer (1974), O’Reilly and New (1982), Lo et al. (1984), Mair et al. (1993) and Bracegirdle and Mair (1996). On the other hand, surface loading will also determine the amount of load acting on the tunnel in addition to overburden when considering a tunnel built on shallow ground. Internal tunnel support has to be erected upon consideration of surface loads and overburden ground loads.
Construction methods
There are many methods available for tunneling in soft and weak ground such as New Austrian Tunneling Method (NATM), tunnel boring machines (TBM), cut and cover method, shield with concrete or segmental linings, micro tunneling technique for small diameter tunnels etc. The most advanced tunneling methods used in urban area are the bentonite slurry balance shield and the earth pressure balance shield (EPB). These methods have revolutionized the conventional methods of tunneling and had replaced the cut and cover method although the latter method is still used in certain cases. Fujita (1989) featured the construction of the tunnel for the Metropolitan Expressway interconnecting Tokyo and Yokohama which was built using the cut and cover method. The construction produced large settlement and caused displacement of the nearby piers which resulted in great environmental problems to the surrounding ground and buildings. Though the shield was invented by Brunel in 1819 in Britain, the technological advancement in shield tunneling was extensively developed in Japan where the first EPB shield was designed in the early 1960s. The EPB and the bentonite slurry shield could minimize the ground settlement greatly to produce a smooth excavation although health hazards were reported by workers due to the high pressured working condition in the enclosed shield which could cause respiratory problems and other defects. The working principle of the EPB shield is in balancing the pressure of the soil inside the chamber with the earth pressure and water pressure acting on the face. The bentonite slurry shield machine works in the same manner except that the earth pressure is used to balance the overburden pressure and the tunnel face is acted solely by the ground water pressure. Although the EPB and bentonite slurry shields are used extensively in urban tunnel construction, their applicability is limited when hard materials or unusual soils are encountered or when large diameter tunnels are required as shown in the construction of the Athens Metro which is underlain by Athenian schists, which consists of a complex mixture of sandstones, limestones, marls, siltstones and slates (Kavvadas et al.,1996). Thus, the construction process is divided into NATM (or shotcreting) and using TBM shield for different segments of the tunnel. Clough and Leca (1993) described the use of two EPB shields for tunneling in a mixed face of sand or sand and gravel overlying clay. Farias and Assis (1996) had also described a case study of tunnel construction in collapsible residual soil whereby NATM was used instead of slurry or EPB shields. When these situations occur, conventional methods of mechanical shield machines or excavator machines are used. Such inconsistencies in tunnel works cause unnecessary wastage of time and resources which will add to extra cost. However, the use of the EPB or bentonite slurry shields is still preferred in difficult and urban grounds for its low cost and risk. Table 2 shows some examples of past projects that involve EPB, slurry shields, NATM, and conventional hand excavation.
Tunnel support system
Tunnel construction has gone through major advancement in excavation techniques over the years with the use of hi-tech machineries, ground reinforcement or ground improvement to stabilize the tunnel face and prevent any tunnel collapse. The advance rate of tunnel excavation has consequently increased significantly with the introduction of the tunnel boring machines (TBM) and newer tunneling methods like the New Austrian Tunneling Method (NATM) or using Earth Pressure Balanced shield (EPB) and slurry shield to stabilize the tunnel face during excavation. There exist many types of ground improvement and reinforcement methods to stabilize the cavity and every individual method will produce different action on the cavity or tunnel face to effect the necessary stabilization. Lunardi (2000) conducted field investigations of Italian tunnels excavated in rocks and subsequently, he categorized these support methods into three different groups whereby each group would exert a different kind of effect on the cavity.
Table 2. Different tunneling projects involved in EPB, NATM
Slurry shields and cut and cover method.
In his study, he found that the rigidity of the core will determine the stability of a tunnel since the deformation of the advanced core causes the extrusion of the face, pre convergence behind the face and the convergence of the cavity as shown in Figure 6. Table 3 shows the categorization of the different kinds of stabilization instruments. The use of ground reinforcement to maintain the stability of the tunnel face during excavation will usually reduce the problem of the weakening of the ground that will cause tunnel collapse and consequently, produce disturbance to existing surface structure. The problem of tunnel collapse will cause the land above the tunnel to subside especially when the excavation is carried out at shallow depth and in weak and soft soils.
Figure 6. Types of deformation in tunnel excavation (longitudinal cross sectional view).
Table 3. Kinds of stabilization instruments and the type of effect exerted (reproduced from Lunardi, 2000)
Figure 7 shows a schematic diagram of the conventional support system for a large scale excavation. The pre-support methods such as steel pipe umbrella, jet grouting, pipe- roof, mechanical pre-cutting and sub horizontal fiber glass face reinforcement, whereby reinforcements are installed prior to excavations, are widely used in large scale excavations in urban areas where subsidence have to be limited. In particular the pre-support capabilities of the umbrella method, which consists of the steel pipe umbrella arch, jet grouting and pipe roof method, have gained popularity and are shown in Figure 8. Numerous field cases of tunnels excavated using the umbrella method were reported by Barisone et al. (1982),Gangale et al. (1992), Borchi et al. (1992), Murata et al. (1996), Shin et al. (1999), Haruyama et al. (2001) and Sekimoto et al. (2001).
Figure 7. Schematic diagram of typical tunnel support system.
Figure 8. Steel pipe Umbrella, Pipe Roof and Jet grouting method.
Although the umbrella method is popular and widely used, one of the difficulties facing tunnel designers is that there are no simple approximations to simulate these methods in numerical analyses. Most researchers tend to crudely approximate the reinforced umbrella arch as a composite material with a combined strength. These crude approximations tend to produce inaccurate surface settlement results. However, Tan and Ranjith (2003) modeled the steel pipes individually for a test model to study the behavior of the ground excavated in soft ground using pipe roof method as a pre-support. Their study shows that surface settlement can be significantly reduced by using a pre-support method. However, further research should be focused on producing simple but reliable and accurate approximations for tunnel support system.
Environmental impact
Health hazards for tunnel workers due to the working conditions in high pressured shields, collapsing of the tunnel face, cave-ins and flooding of tunnels are important factors to be considered before undergoing any tunnel constructions. Surface subsidence also causes safety concerns for the public because of roadway cave-ins or structures and buildings tilting or even collapsing. Services carrying water, gas or sewage pipes are also vulnerable to ground deformation caused by tunneling. Studies have been made by Bracegirdle and Mair (1996) where they looked into the effect of tunneling on nearby services and described a methodology for evaluating the potential damage to cast iron pipes where settlements were due to tunneling in London clay. Mair et al (1993) derived a value of i at the depth of the services, (zp) to estimate settlements for tunnels in clay.
![]() |
(12) |
Attewell et al. (1986) also gave comprehensive analysis on strain field calculations for pipes lying transverse or parallel to the tunnel axis due to ground movements induced by tunneling. However, Yoshida and Kusabuka (1994) discovered that existing services and tunnels were influenced more significantly by the additional earth pressure from shield tunneling than by the subsidence due to stress release. They also proposed that the displacement and deformation of the existing services and tunnels are affected primarily by the ratio of the cross section area of the underground pipe and the cross section area of the shield, and the distance between the shield and the underground pipe.
CONCLUSION
Although empirical methods of predicting soil settlement due to tunneling induced ground movements have been used extensively and successfully over the years, a theoretical solution is still relevant and important considering variations in site conditions and soil ambiguity. In addition, tunneling methods and tunnel support systems also differs depending on site conditions and tunnel geometry and hence empirical solutions have their limitations. A comprehensive theoretical solution will require primarily the consideration of soil characteristics, tunnel geometry, site investigations, tunnel support systems and construction methods. This solution should be substantiated by calibrating with field case histories and 3-D numerical analyses.
references
![]() | |
© 2003 ejge |