A discrepancy between observed and predicted NATM tunnel behaviors and updating: a case study of the Sabzkuh tunnel
Abstract
The 11 km long Sabzkuh water conveyance tunnel is under construction using conventional and tunnel boring machine tunneling techniques in the Zagros Mountains of south west Iran. During excavation of the conventional section (i.e., 350 m), three daylighting collapses occurred in the tunnel, and large sinkholes developed at the ground surface, due to both poor ground conditions and inappropriate selection of excavation and support class (ESC). In this study, the influence of ESC on tunnel response and stability was investigated through three dimensional (3D) finite difference and finite element methods. The analysis results of numerical modeling indicated that the tunnel face would be stable if a single stage (full face) excavation and an advance step of 1.5 m (defined as ESC#1) were applied. However, the reality was different, and the tunnel collapsed soon after the face was only 35 m away from the portal. Having changed the ESC by taking into account new geological and geotechnical data obtained from face mapping, a field survey of surface cavity, and a trench study, the tunneling advance rate was satisfactory (1.5 m/day) despite the many challenges encountered. This paper presents a brief review of some of the key geological challenges faced in the tunnel design, including characterization of the ground conditions, selection of appropriate design parameters, and evaluation of the excavation and support installation sequence based on monitoring and analyzing ground behavior during construction. The discrepancy between the predicted and observed behavior of the tunnel, and some practical considerations, are also discussed.
Keywords
Sequential excavation method Fault zone Daylighting collapse Sabzkuh tunnel Numerical methodIntroduction
Tunneling in soft ground conditions is always challenging due to the various geotechnical and environmental constraints encountered. Some new findings and developments in the field of soft ground tunneling have emerged in recent decades to address these problems. Most of these challenges are related to geological and geotechnical conditions. Geotechnical aspects of soft ground tunneling have been addressed by many researchers (e.g., Sterpi et al. 2013; Elyasi et al. 2015). However, obtaining relevant information on tunnel ground conditions has been the key factor for successful construction of underground structures in difficult ground conditions.
Tunneling in the young Zagros Mountains in south-western Iran has always posed challenges due to existing geological structures such as faults and severe natural folds, as well as diverse stratigraphic and active tectonic zones (Shahidi and Vafaeian 2005). Hence, performing a detailed site investigation has become one of the most important prerequisites in the design and construction of tunnels in this region. The Sabzkuh–Choghakhor water conveyance tunnel is currently under construction in this region. Three daylight collapses have occurred during the construction of the tunnel, forcing changes to the excavation and support class (ESC) for the rest of the tunnel.
Here, we present the problems that occurred in the first 390 m of Sabzkuh tunnel during its construction, emphasizing the importance of an appropriate ESC, as well as monitoring and computational tools based on numerical analyses and practical considerations.
Project overview and geology
The tunnel is going to be excavated using a double-shield tunnel-boring machine (DS TBM). Information obtained from the site, including field reconnaissance, geotechnical and geophysical studies, revealed that the tunnel, from the beginning to chainage 0 + 390 km (section T1), is located in alluvium. Taking into consideration the application range of DS TBM, and its limitation for working in alluvial soil, it was decided to excavate this part of the tunnel using conventional methods (Karami et al. 2014; Eftekhari et al. 2014).
Using the supplementary geotechnical data, the consultant engineers were able to review the tunnel construction process and update the design by using additional data collected during the construction process. However, in the rest of the tunnel, various challenges, including active tectonics, existence of fault zones, poorly consolidated and cohesionless ground conditions, and confined and non-confined underground aquifers, were expected.
Stability analysis and design of the tunnel support system
For stability analysis, and with the aim of achieving an appropriate tunnel model, several analytical, observational and empirical methods are often used in parallel. In recent years, numerical modeling using the finite element method (FEM) (Hajjar et al. 2014; Sharifzadeh et al. 2013), the boundary element method (BEM), the distinct element method (DEM) and the finite difference method (FDM) (Vlachopoulos and Diederichs 2014; Elyasi et al. 2015) has been applied widely in rock engineering problems (Jing and Hudson 2002; Fairhurst and Carranza-Torres 2002; Hoek et al. 2008; Carranza-Torres et al. 2013). By using these methods, modeling of complicated geometry under different loading conditions and employing appropriate constitutive models is possible.
Estimation of loads on support system
In his book Theoretical soil mechanics, Terzaghi (1943) presented one of the first models for computing the stability of sections of shallow tunnels excavated in soils, based on the principle of soil arching. This model is based on the condition that the soil extending laterally and above the tunnel is in a state of limit equilibrium. A method for computing support requirements for (both shallow and deep) tunnels in both soils and rock was later developed from the Terzaghi original model. This became to be known as the Terzaghi’s ‘rock load’ method (Proctor and White 1946) and has been a popular method for designing tunnel supports in many countries since its introduction (Carranza-Torres et al. 2013).
Estimation of support load in Sabzkuh tunnel, section T1
Load | Author | Value (MPa) | Equation |
---|---|---|---|
Vertical load | Terzaghi | 0.136 | \( \frac{{B_{1} \left( {\gamma - \frac{2c}{{B_{1} }}} \right)}}{k\tan \phi }\left[ {1 - e^{{\frac{KD\tan \phi }{{B_{1} }}}} } \right] \) |
Bierbäumer’s theory | 0.287 | \( \gamma H\left[ {1 - \frac{{\tan \phi \cdot \tan^{2} (45 - \phi /2)H}}{b + 2m \cdot \tan (45 - \phi /2)}} \right] \) | |
Balla’s theory | 0.177 | \( \gamma H\left[ {F_{H} + \frac{b}{H}F_{B} - \frac{c}{\gamma H}F_{c} } \right] \) | |
Analyses | 0.2 | ||
Horizontal load | Terzaghi | 0.046 | 0.3γ(0.5 m + h_{p}) |
0.059 | \( \frac{{B_{1} \left( {\gamma - \frac{2c}{{B_{1} }}} \right)}}{k\tan \phi }\left[ {1 - e^{{\frac{KD\tan \phi }{{B_{1} }}}} } \right] \) | ||
Rankin theory | 0.258 | \( \gamma H\tan^{2} (45 - \phi /2) \) | |
Analyses | 0.09 |
Meanwhile, for safety reasons, the lining horizontal and vertical pressures in Sabzkuh tunnel were set at 0.09 MPa and 0.2 MPa, respectively.
Analysis of a composite liner with the ‘equivalent section’ approach
Input data for the analysis of a semi-circular liner comprised of shotcrete and steel sets according to the ‘equivalent section’ approach from Carranza-Torres and Diederichs (2009)
Excavation and support class#1
Soil properties as input data in the finite difference method (FDM) model
Parameter | Unit | Value |
---|---|---|
Type of material behaviour | – | Drained |
Total unit weight | kN m^{−3} | 18.50 |
Young’s modulus | MPa | 40 |
Poisson’s ratio | – | 0.25 |
Cohesion | kN m^{−2} | 35.0 |
Friction angle | ° | 25.0 |
Excavation and support class#1 (ESC#1) for the tunnel
Analysis of the collapse factors and designing the construction after the tunnel daylight collapse
Taking into account the inefficiency of the existing excavation method, a database was created using experiences from similar projects and data obtained from field studies and studies conducted during construction in this region. Moreover, it was decided to conduct a complementary site investigation to determine the link between tunnel collapse and the geological settings in this section of the tunnel.
The results of the review indicate that a complex geological model of the region requires the continuous identification, analysis and monitoring of the movement of the surrounding soil mass to achieve the most economic and stable excavation, and to support the methods chosen in different conditions of the tunnel.
Geophysical investigations were carried out with different geophysical methods such as common reflection point (CPR) and reflection seismology in the first 3 km of the tunnel route. Several faults were identified in the portal area and alluvial section. In addition, multiple soil samples were taken from different soil layers at the portal and the geological profile for the area under study was modified (see Fig. 2).
Geotechnical properties of soil layers at the tunnel portal
No. | Soil unit | γ (g/cm^{3}) | φ (°) | C (kg cm^{−2}) |
---|---|---|---|---|
1 | CL | 1.71 | 22 | 0.49 |
2 | CH | 1.64 | 19 | 1.1 |
3 | ML | 1.76 | 26.7 | 0.18 |
4 | CL-ML | 1.73 | 24 | 0.35 |
5 | SC-SM | 1.88 | 31.3 | 0.035 |
6 | GP-GM | 2.1 | 34 | 0.01 |
7 | Slide surface | 1.8 | 65 | 0.0 |
Accordingly, a series of 3D numerical modeling using the FEMs were performed in order to investigate the influence of active span and partial-face excavation on tunnel stability.
Using the developed database, numerical modeling and continuous monitoring of the tunnel behavior, ESCs were selected for different geological units along the tunnel alignment and remedial solution in collapse-prone zones.
Excavation and support class#2
Taking into consideration the ineffectiveness of the full face excavation method as a construction solution, the ground behavior and classification were re-examined to define a suitable excavation method and construction process. Accordingly, an excavation scheme was adopted considering different factors such as the tunnel section, soil type, ground behavior and classification, and underground water table (Sharifzadeh et al. 2013). Taking all these factors into consideration, the sequential excavation method (SEM) was selected as the most appropriate method with which to continue the construction.
FEM analyses, in conjunction with empirical methods, were used to evaluate potential ground behavior upon tunnel excavation, and to determine the required excavation sequence and support measures.
Stability analysis for sequential excavation method
In this study, 3D finite element models with the capability of simulating sequential excavation were adopted in order to simulate the construction process in the Sabzkuh tunnel; each construction stage may involve soil excavation and/or support installation.
Material properties of the soil
Parameter | Symbol | Unit | Description/value |
---|---|---|---|
Material model | – | – | Mohr–Coulomb |
Type of material behavior | – | – | Drained |
Soil weight above phr. level | γ_{unsat} | kN m^{−3} | 15.0 |
Soil weight below phr. level | γ_{sat} | kN m^{−3} | 18.5 |
Young’s modulus (constant) | E_{ref} | kN m^{−2} | 5 × 10^{4} |
Poisson’s ratio | ν | – | 0.30 |
Cohesion (constant) | c_{ref} | kN m^{−2} | 17.0 |
Friction angle | ϕ | ° | 28.0 |
Dilatancy angle | Ψ | ° | 0.0 |
Excavation and support class#2 (ESC#2)
Monitoring
Ground deformation monitoring in tunnelling is a common means of selecting and controlling the excavation and support methods among those predicted in design, ensuring safety during tunnel excavation (including the safety of personnel inside the tunnel and structures at the ground surface) and construction quality (Kavvadas 2005).
Tape extensometer technical specifications
Parameter | Unit |
---|---|
Measurement | mm (in.) |
Available ranges | 20, 30 m and 66, 100 ft |
Measurement range | 50 mm |
Accuracy | 0.01 mm |
Repeatability | 0.1 mm |
Weight | 2.2 kg |
The monitoring observations were in a good agreement with the results obtained from numerical analyses. However, on the right wall of the tunnel, due to high groundwater condition, larger deformations were observed in comparison with the left side. This problem was resolved by installing drainage facilities in this area.
Excavation and support class#3
According to the experiences gained from excavation in the first section of the Sabzkuh tunnel and surface and the subterranean studies conducted, the ground conditions in this section indicate a complex geological model, along with numerous uncertainties largely pertaining to soil characteristics, fault zones, pore water pressure, sliding surfaces, alternation and dip angle of soil layers.
Therefore, it was essential to use ground improvement techniques such as pre-grouting, and also predict specific measures to achieve a safe construction process. The collapse occurred at a distance of 150 m and ground adverse conditions (i.e., a fault zone with a length of 50 m) made the rest of the excavation problematic without pre-reinforcement systems.
Excavation and support class#3 (ESC#3)
The results proved to be highly satisfactory from both technical and economic points of view. Given the unfavorable ground conditions, the presence of groundwater, and an advance rate of 1 m/day (contrary to 1.5 m/day for ESC#2), a good and regular plan for construction process and an acceptable level of safety for tunnel operation have been achieved. In addition, analysis of results obtained from the monitoring process showed that, with implementation of the new pre-support system, the ground strength surrounding the tunnel increased, and displacement and pressure imposed on support systems decreased compared to ESC#2.
Concluding remarks
Due to the complicated geological conditions, tunneling in the Zagros Mountains has been always challenging, with consequent time delays and cost overruns in project completion. In the construction process of the Sabzkuh tunnel, a geotechnical model was planned, implemented and analyzed independently and regardless of the geological model. The lack of a logical relationship between these two models, as well as the complex ground conditions, together with other uncertainties, provided inappropriate information to the designers, affecting the analytical model as well as the excavation method. This led to an early daylighting collapse when the tunnel face reached an advance of 35 m from the tunnel portal. Unstable ground and collapsing zones, and the required techniques for passing through were the on-going challenges during construction of the tunnel. The ground conditions and scale of the collapse revealed that the tunnel goes through a complicated geological model, which makes the reconnaissance, analysis and continuous monitoring of ground movement inevitable. Accordingly, the single stage (full face) excavation was shifted to a NATM/SEM method due to its applicability and flexibility when confronted with the actual tunnel site conditions. A data server was established in order to gather data from field studies and as-built findings, regarding previous experiences. This server helps engineers review the plan and improve the design for the remaining tunnel sections.
Experiences obtained during the construction of the Sabzkuh tunnel in the soil (T1) section revealed that the complicated geological model requires frequent exploration, analysis and behavior checking of the surrounding soil mass in order to find the most economical and stable excavation model and support system for different conditions encountered along the tunnel alignment. Hence, the project owners should take into consideration preventive measures and the most practical and feasible solutions with an appropriate factor of safety according to the changing ground conditions encountered during the construction process of this tunnel.
This case study demonstrates once again the importance of early detection or prediction of potentially problematic zones, for instance, through probe drilling and monitoring in tunneling, especially in changing or difficult ground conditions characterized by alternating layers, faulting and localized zones of high water pressure.
Notes
Acknowledgments
Many thanks are due to IMN Consulting Engineers and the project manager of the Sabzkuh tunnel, Mr. Ali Asghar Izadi for sharing their experience and guidance with the authors.
References
- American Society of Civil Engineering (1984) Guidelines for tunnel lining design. In: O’Rourke TD (ed) ASCE technical council on research. Technical Committee on Tunnel Lining Design. American Society of Civil Engineers, New YorkGoogle Scholar
- Carranza-Torres C, Diederichs M (2009) Mechanical analysis of circular liners with particular reference to composite supports. For example, liners consisting of shotcrete and steel sets. Tunn Undergr Space Technol 24(5):506–532CrossRefGoogle Scholar
- Carranza-Torres C, Reich T, Saftner D (2013) Stability of shallow circular tunnels in soils using analytical and numerical models. In: Proceedings of the 61st Minnesota annual geotechnical engineering conference, University of MinnesotaGoogle Scholar
- Eftekhari A, Taromi M, Saeidi M (2014) Uncertainties and complexities of the geological model in slope stability: a case study of Sabzkuh tunnel. Int J Min Geo-Eng 48(1):69–79Google Scholar
- Eftekhari A, Taromi M, Saeidi M (2015) A study of the reinforcement effect of IPE Arc Support Technique (IAST)—a case study of Sabzkuh tunnel. J Eng Geol 9(1):2559–2728Google Scholar
- Elyasi A, Javadi M, Moradi T, Moharrami J, Parnian S, Amra M (2015) Numerical modeling of an umbrella arch as a pre-support system in difficult geological conditions: a case study. Bull Eng Geol Environ 75:211–221Google Scholar
- Fairhurst C, Carranza-Torres C (2002) Closing the circle. In: Labuz J, Bentler J (eds) Proceedings of the 50th annual Geotechnical Engineering Conference, MinnesotaGoogle Scholar
- Hajjar M, Nemati Hayati A, Ahmadi MM, Sadrnejad SA (2014) Longitudinal settlement profile in shallow tunnels in drained conditions. Int J Geomech 15(6):04014097CrossRefGoogle Scholar
- Hoek E, Carranza-Torres C, Diederichs M, Corkum B (2008) Integration of geotechnical and structural design in tunneling. In: Proceedings of 56th Annual Geotechnical Engineering Conference, MinneapolisGoogle Scholar
- Itasca Consulting Group (1997) FLAC3D, fast Lagrangian analysis of continua in 3 dimensions. Version 2.1. MinneapolisGoogle Scholar
- Jing L, Hudson JA (2002) Numerical methods in rock mechanics. Int J Rock Mech Min Sci 39(4):409–427CrossRefGoogle Scholar
- Karami M, Faramarzi L, Bagherpur R, Raisi Gahrooee D (2014) Influence of geological features and geomechanical properties of rock mass on TBM selection for Sabzkuh water conveyance tunnel. J Eng Geol 8(2):2169–2198Google Scholar
- Kavvadas M (2005) Monitoring ground deformation in tunneling: current practice in transportation tunnels. Eng Geol 79(1–2):93–113CrossRefGoogle Scholar
- Oke J, Vlachopoulos N, Marinos V (2014) Umbrella arch nomenclature and selection methodology for temporary support systems for the design and construction of tunnels. Geotech Geol Eng 32:97–130CrossRefGoogle Scholar
- Proctor RV, White TL (1946) Rock tunneling with steel supports. Commercial Shearing, OhioGoogle Scholar
- Shahidi AR, Vafaeian M (2005) Analysis of longitudinal profile of the tunnels in the active faulted zone and designing the flexible lining (for Koohrang-III tunnel). Tunn Undergr Space Technol 20:213–221CrossRefGoogle Scholar
- Sharifzadeh M, Kolivand F, Gorbani M, Yasrobi Sh (2013) Design of sequential excavation method for large span urban tunnels in soft ground—Niayesh tunnel. Tunn Undergr Space Technol 35:178–188CrossRefGoogle Scholar
- Sterpi D, Rizzo F, Renda D, Aguglia F, Zenti CL (2013) Soil nailing at the tunnel face in difficult conditions: a case study. Tunn Undergr Space Technol 38:129–139CrossRefGoogle Scholar
- Terzaghi K (1943) Theoretical soil mechanics. Wiley, New YorkCrossRefGoogle Scholar
- Vlachopoulos N, Diederichs MS (2014) Appropriate uses and practical limitations of 2D numerical analysis of tunnels and tunnel support response. Geotech Geol Eng 32:469–488CrossRefGoogle Scholar