Parametric studies on prediction methods of face support pressure and surface settlement for soft ground tunneling with epb tbms

Abstract

Throughout history, tunnelling has been done for war, mining, transportation and various other purposes. Nowadays due to the rapid growth of technology and industry, the population is increasing rapidly especially in cities where have limited space. In recent years there has been a big necessity for new infrastructure systems in urban areas because of rapid increase of population. Highway tunnels, metro tunnels, water tunnels are the most common infrastructure system types in tunnelling application areas. The number of tunnels excavated for transportation purpose has been increasing rapidly in urban areas all around the world. Tunneling demand in urban also requires high-quality work in limited space at limited times. The methods of underground excavations have changed with the development of technology. By using developed Tunnel Boring Machines (TBMs) which enable full face excavation, it is possible to make faster and quieter excavations with lower vibration compared to the traditional methods. In this way, tunneling induced surface settlement is minimized. One of the most important problems in urban tunnelling is the displacements occurring in the surrounding buildings (surface) due to the excavation. For this reason, most of the urban tunneling is performed with soft ground TBMs. Since most of the tunnels are shallow in urban areas and soil is mostly poor quality, effects of underground excavation may reach to surface easily. Especially in this type of soil, applying correct face pressure is one of the most significant factors in excavations carried out with soft ground TBMs. Therefore, it is very important to determine the face support pressure to be applied, and estimate the surface settlement amount accordingly for safety of tunnel and not to damage the structures on surface. There are empirical, analytical and numerical methods to predict both face support pressure and surface settlements in feasibility stage of a tunnel project. These empirical and analytical methods have some restrictions, and their applicability may vary depending on excavation methods, soil properties, tunnel diameters and overburden height. In such cases, the numerical methods having less restrictions are more appropriate, since more input parameters are enabled in numerical methods. In particular, finite difference and finite element methods are among the most widely used numerical methods in recent years. It is stated in many studies that the numerical analysis give more realistic results compared to analytical and empirical methods. The aim of this thesis study is to investigate the methods developed for predicting both face support pressure and surface settlement with parametric studies. Methods used to calculate face support pressure in this study are: (i) empirical methods based on empirical formulas, mostly derived from limited previous tunnel case studies and field observation and measurements (Broms and Bennemark, 1967; Davis et al. 1980; Kimura and Mair, 1981), (ii) analytical (theoretical and/or semi-theoretical) methods (Atkinson and Potts, 1977; Krause, 1987; Leca and Dormieux, 1990; Jancsecz and Steiner, 1994; ITA 2000; Carranza-Torres, 2004) and (iii) numerical methods (finite element method, RS2 software). Methods used to calculate maximum undrained surface settlement in this study are: (i) empirical methods (Arıoğlu & Schmidt (after Çopur et al. 2007); Herzog (after Schmidt, 1985), (ii) analytical (theoretical and/or semi-theoretical) methods (Limanov, 1957) and (iii) numerical methods (finite element method, RS2 software). The results obtained from these empirical, analytical and numerical methods are compared. At the same time, it is examined that how sensitive these methods are to the change of parameters within themselves. The results obtained are intended to guide future studies. Numerical models used in this study are created by RS2 two dimensional software using Finite Element Method (FEM). There are two types of analysis in RS2 which are axisymmetrical and plain strain analysis. An axisymmetric analysis allows that analyze a 3-dimensional model which is rotationally symmetric about an axis. Although the input is 2-dimensional, the analysis results apply to the 3-dimensional problem. While axisymmetric analysis is used to predict face support pressure, plane strain analysis is used to predict surface settlement in this study. The reference model and variables to be used in parametric studies are first determined. Some of the methods proposed to determine face support pressure are valid for cohesive soils while some methods are valid for granular soils. For this reason, geotechnical properties of reference models are assumed by taking the literature into consideration. Two different soil models such as granular and cohesive are created for face support pressure analyzes. Within the scope of this study, investigated soil parameters are internal friction angle, cohesion, unit weight, Young's modulus and Poisson's ratio. Tunnel diameter, tunnel depth (overburden) parameters, and thus the ratio of depth to diameter, are technical parameters used in the study. Another reference model is created to use in parametric studies on surface settlements. Same soil parameters such as internal friction angle, cohesion, unit weight, Young's modulus and Poisson's ratio and technical parameters which are tunnel diameter and tunnel depth are investigated on parametric studies to calculate maximum undrained surface settlement in this study. In all parametric studies, variables are changed in their value ranges and results are obtained for each approach. The results obtained are firstly evaluated within themselves, and then these results are compared with similar studies in the literature. A number of methods which include parameters mentioned above have been developed for both face support pressure and surface settlements so far. In this study, only some of these are investigated in parametric studies. Main results of the study can be summarized as below: • It is not possible to apply some empirical and analytical methods to every project and it is necessary to pay attention to sensitivity of parameters within the used methods. • The method of Broms & Bennemark (1967) includes a number of parameters but it is very important to determine the cohesion value correctly in this method. • Using the ITA (2000) method in cohesive soils may be misleading. • Krause (1987) and Carranza-Torres (2004) methods do not contain tunnel depth parameter while this is the most influential parameter in other methods. Therefore, there is a concern about the reliability of these methods. • Result of parametric studies belong to Krause (1987) give the lowest results in analysis of cohesive soil. Obtained values of face support pressure for granular soil calculations from method of Leca & Dormieux (1990) are compatible with method of Jancsecz & Steiner (1994). Results of these two methods have the lowest values in calculations for granular soil. • Herzog (1985) method used for estimation of maximum surface settlement gives higher results compared to the other methods. • Information of only increasing or decreasing C/D ratio is not sufficient for making an inference to determine its effect on surface settlement or face support pressure. The result may vary depending on which parameter changes. • Tunnel depth, internal friction angle and cohesion of soil are the most important parameters in calculating face support pressure considering all methods examined in this study. On the other hand tunnel depth and tunnel diameter are the most effective parameters for analytical and empirical methods to determine maximum surface settlement. • Especially using of sensitive parameters incorrectly in estimations may cause results to be far from the actual field values. It should be emphasized that these parametric studies are performed with parameters assumed. In particular, the sensitivity of the methods to the varying parameters are directly related to the accepted range of the variables. There are other parameters such as surface surcharges, underground water condition, having different geological layers that are not investigated in this study. It should be noted that different results can be obtained by taking these parameters into account.