Application of Jet Grouting in NATM Tunneling

Ch. Brandstätter mail@fest.tuwien.ac.at; R. Lackner; Ch. Pichler; H.A. Mang    Institute for Strength of Materials, Vienna University of Technology Karlsplatz 13, A - 1040 Vienna, Austria, Europe

INTRODUCTION

In order to meet the continuously increasing traffic in urban areas, subterranean constructions have gained considerable importance. With respect to above-ground transportation systems, they, exhihit remarkable advantages, such as (see, e.g., Golser et al. (1997))

• little interference with the existing overground infrastructure,

• greater flexibility concerning the design of transportation systems;,

• minimization of the required above-ground area,

• minimization of the disturbance of human beings resulting from traffic noise,

• and, as a direct consequence of the last two items, relatively wide acceptance of such infrastructural means by the population.

For the construction of tunnels in urban areas, the New Austrian Tunneling Method (NATM) has proved to be well suited. When driving tunnels according to the NATM, after the excavation of a cross-section of a tunnel, shotcrete is applied onto the tunnel walls, constituting a thin and flexible shell. The NATM is characterized by a strong interaction of the hardening/creeping shotcrete shell and the viscous soil and/or rock, exerting pressure on the lining. The creep properties of shotcrete are the source of the deformations required for the activation of the load-carrying capacity of the surrounding rock and soil formation. In urban areas, however, these deformations must be minimized in order to avoid damage of buildings and infrastructure. In these cases, the soil surrounding the tunnel is improved by means of grouting techniques or soil freezing. More recently, horizontal jet grouting (HJG) is used for this purpose. According to its designation, HJG is performed horizontally at the tunnel face. Applications of HJG are reported in (Gioda and Locatelli, 1999) (Guatteri et al., 1994) (Fitl, 1997) (Liebsch and Haberland, 1998).

The advantage of jet grouting is its applicability to a wide range of different soils. During jet grouting, a thin high-pressure jet of cement grout is discharged laterally into a borehole wall (see Figure 1). This jet mixes with the soil resulting in a column of modified soil, referred to as soilcrete, with low permeability and increased strength. The radius of this column depends on several factors, such as (Henn, 1996)

• work pressure,

• injection time, which is determined by the rate at which the drilling rods are rotated and withdrawn (see Figure 1(b)), and

• the properties of the in situ soil.

The improvement of the material properties by means of jet grouting mainly depends on the amount of cement in soilcrete. In general, this amount is unknown. Moreover, no appropriate model for determination of the radius of soilcrete columns exists. Still, test columns are required in order to obtain the radius for a certain soil and a certain set of grouting parameters. Hence, as regards numerical analyses of HJG in the context of NATM tunneling, the following questions arise:

• How do soil conditions and production parameters influence the dimensions of the soilcrete body and the amount of cement of soilcrete?

• What are the properties of the material obtained from jet grouting, referred to as soilcrete?

• How does soil improvement by means of HJG change the deformations during tunnel excavation?

DETERMINATION OF PROPERTIES OF SOILCRETE BY MEANS OF BACK ANALYSIS

Recently, a novel strategy for the evaluation of properties of soilcrete was developed at Vienna University of Technology (Lackner et al., 2001a) (Brandstätter et al., 2001). It exploits the exothermal character of the hydration process. Based on in situ temperature measurements and thermochemical analyses both the cement content of soilcrete and the radius of the soilcrete column are computed by means of back analysis. The material model employed for the thermochemical analysis (Ulm and Coussy, 1995) is restricted to weak couplings such that the evolution of the hydration reaction only depends on the mass of hydrates formed and on the current temperature. The employed evolution law for the degree of hydration ξ of Arrhenius type reflects the thermally activated nature of the chemical reaction. It is given as (Ulm and Coussy, 1995)

˙=A˜ξexp−EaRT,

where Ea is the activation energy and R is the universal constant for ideal gases (Ea/R ≈ 4000 K, see Bažant (1988)). The chemical affinity à is the driving thermodynamic force of the hydration reaction. T is the temperature in Kelvin.

The field equation for the thermochemical problem is derived from the first law of thermodynamics. In the absence of volume heat sources and of negligible terms, this law is given as (Ulm and Coussy, 1995)

cT˙−ℓξξ˙=−divq,

with ρc[kJ/(m3K)] as the heat capacity, ℓξ [kJ/m3] as the latent heat of hydration per unit volume, and q as the heat flow vector, q is related to the temperature T via Fourier’s linear (isotropic) heat conduction law,

=−kgradT,

with k [kJ/(m h K)] as the thermal conductivity.

For the thermochemical analysis of a soilcrete column, an axisymmetric model is employed (see Figure 2). The developed strategy was applied to a test column at a Vienna construction site. Input for determination of the radius of the soilcrete column, R, and the cement content of soilcrete are in situ temperature measurements at the center of a soilcrete column (see Figure 3(b)). The temperature sensor was installed after jet grouting at a depth of 2 m. The height of the soilcrete column is 3 m, starting 0.5 m below ground level, see Figure 3(a). Back analysis was performed in an iterative manner: First, reasonable values for the radius of action and the cement content were chosen. Based on these values, a thermochemical analysis was performed. In general, the numerically obtained evolution of the temperature does not match the in situ temperature history. Hence, the parameters are adapted and the thermochemical analysis is repeated. Back analysis is terminated if the difference between the in situ temperature history and the respective numerical result is acceptable. 0.8 m and 5 % were obtained from back analysis as the final values for the radius and the cement content, respectively. The respective numerical result is shown in Figure 4(b). Good agreement between the determined and actual radius of the soilcrete column (see Figure 4(a)) is observed.