This paper analyses a tunnel length of 850m of S. Emiliano formation, in Pajares tunnels. The main objectives are the recapitulation of problems occurred during tunneling works and the proposal of technical solutions, based on the experience gained during the works. Moreover, the RME index has been applied to the analyzed stretch. This index predicts TBM average rates of advance, ARA. The results show that the actual advance rates obtained during the excavation differ from theoretical values, mainly due to TBM problems and to the weak rock mass. The selection of a specific TBM determines the works’ future development, thus requiring exhaustive geological studies. It also confirms that for tunnels currently being excavated, longer than the existing ones, ground treatments will be systematically needed. These ground treatments imply significant difficulties, especially when performed from the TBM. Thus, TBMs must be improved and adapted to deal with future challenges.
1.1 Description of Pajares tunnels
Pajares Tunnels, two parallel tubes each 24.6 km in length, were excavated with five TBMs in six different working faces as part of the León-Gijón section of the Spanish high-speed rail network. The works were divided in four sections. Two single shield hard rock TBMs (one from Herrenknecht and one from NFM– Wirth) were used for Section 1. A third machine, a double shield hard rock Herrenknecht TBM, worked on Section 2 (consisting of a 5.5-km-long access tunnel and the 4.5-km-long twin of the central part). Two single shield TBMs (NFM-Wirth and MHI-Duro Felguera-Robbins) worked on Sections 3 and 4, each of 10.3 km in length.
The section of the tunnels is circular of approximately 52m², with an inner diameter of 8.5 m. The support was formed by a ring of precast segments of 0.5m thick, which implies an excavation diameter of 10.10 m.
The maximum expected overburden was 1000 m, surpassing 800m in different parts of the tunnels; this implies a high stress state of the rock mass, which has caused squeezing problems in the analyzed formation.
The duration of the works (from July, 13, 2005 to July, 11, 2009), together with the total length of the tunnels, the intricate geology of the area, the complicated orogeny of the materials and the depth of the tunnels give an idea of the difficulty of the excavation and the challenges faced.
Borehole failure data provides a unique insight in the characteristics of the stress field because in deep boreholes it tends to be pervasive, providing the opportunity to study variability of the stress components. This variability might follow self-affine scaling and could be related to scaling characteristics of the natural fractures network and earthquake magnitude-frequency statistic. If this were the case, then the measurable variations in stress orientation could be used to constrain statistical attributes of the fracture network and to anticipate the seismic response of a rock mass. In this paper, we evaluate seven techniques to determine the fractal dimension, D, of stress orientation variations indicated by wellbore failure data by applying them first to synthetic data of known fractal dimension. Particular attention was given to assess the biases introduced by the presence of gaps and noise in the data. Finally, the evaluation techniques were applied to real borehole failure data from Soultz-sous-Forêts and Basel. Preliminary results indicate that significantly different estimates of D were found for different methods applied to the same dataset which are best explained as reflecting the impact of gaps in the data.
Knowledge of stress distribution is central to rock mass characterization of deep engineering projects. Wellbore failure provides an opportunistic window for defining stress state because at large depth it tends to be pervasive. In such cases it is possible the generate almost continuous profiles of certain attributes of the stress tensor and hence evaluate the nature of the variability that is seen.
There are indications that in-situ stress variations, as many other geological phenomena, might follow self-affine scaling (Turcotte and Huang, 1995). It is hypothesized that such scaling characteristics are intimately related to the scaling characteristics of the fracture network and earthquake magnitude-frequency statistics (Day-Lewis et al., 2010). If this were the case, then the measurable variations in stress orientation could be used to constrain statistical attributes of the fracture network that are difficult to estimate, and to anticipate the seismic response of a rock mass to hydraulic injections.
In order to be of practical value, the scaling characteristics of stress variability must be reliably determined. This paper evaluates various methods for estimating the scaling characteristics of stress variations by applying them on synthetic series of known fractal dimension. We also evaluate the ability of the methods to deal with gaps in the record and measurement inaccuracy, which is common for borehole failure data. Finally, the various methods are applied to real data sets from three 5 km deep boreholes at Soultz-sous-Forêts and Basel to compare the estimates of fractal dimension of the stress series.
Eurocode 7 is the harmonised European standard dealing with geotechnical engineering design. Although design involving rock masses is included in the code, rock engineering principles do not seem to be fully contemplated. This document summarizes some of the most relevant issues the authors consider that need to be improved in what concerns the application of Eurocode 7 to rock engineering design. This includes topics such as the implications of the discontinuous nature of rock masses, limit states and failure modes, strength criteria, characteristic values and partial factors for rock mass parameters, rock mass characterization, use of classification systems in design, among the most significant issues.
The Structural Eurocodes were conceived as a group of harmonised European standards for the structural and geotechnical design of buildings and civil engineering works, and they are a suite of ten standards concerned with the safety, serviceability and durability of structures. Eurocode 7 (EN1997-1) is one of these standards and is concerned with all aspects of geotechnical design. It deals with constructions in or on the ground, which is defined as “soil, rock and fill in place prior to the execution of the construction works”. Rock engineering design is, therefore, included in the scope of Eurocode 7 (EC7), but this is often overlooked.
The Eurocodes adopt a semi-probabilistic approach of safety verification, based on rules, partially deterministic, that introduce safety at the following levels: selection of appropriate representative values of the various random parameters (actions and resistances); application of partial factors to these parameters; and introduction of safety margins in the various models used in the calculations.
Geotechnical design was slower than structural design in the application of probabilistic or semiprobabilistic approaches. Reasons for this may be the deep roots of empirical methods used in the design of structures in or on the ground. Geotechnical design does not deal with manufactured materials, with relatively well known parameter values, but with natural materials, of a great diversity as regards their origin and the condition in which they are found in nature. Geotechnical structures are not so well defined geometrically as the structure of a building or a bridge, and the actions on them are also more difficult to establish and quantify.
An approximated linearization approach based on the first order reliability method (FORM) is applied to analyse the system reliability of a circular tunnel in a jointed rock mass characterized by the Hoek-Brown non-linear failure criterion. Three failure modes are considered in this paper: exceedance of support deformation capacity, unacceptable tunnel convergence and insufficient rockbolt length. We employ the convergence-confinement method (CCM) to compute the stresses and displacements around the tunnel. Eight parameters—related to rock mass properties, in situ stress and shotcrete properties—are regarded as independent random variables with lognormal distributions. The series system reliability is computed, through the complementary of the intersection of safe domains, based on the results of FORM. An illustrative example is demonstrated with a circular tunnel. We compare the solution of the proposed method with Monte Carlo simulation (MCS) and bimodal bounds. The results show that the proposed approach can be applied to deal, efficiently and accurately, with the system reliability problem of circular rock tunnels.
Reliability methods can deal with uncertainties more efficiently than traditional deterministic approaches. For that reason, given the significant uncertainties that commonly occur in underground construction, reliability-based analysis on tunnel engineering has attracted a significant attention in the last decade (e.g. Kohno et al., 1992; Oreste, 2005; Lü & Low, 2011 etc.). Most previous efforts focused on single-failure modes. But this could be unconservative, as the system probability of failure could be significantly larger than the probability of failure computed for any single mode, especially when failure modes are weakly correlated and have similar probabilities of failure.
Recently, Lü et al. (2013) considered three different limit state functions (LSFs)—exceedance of support deformation capacity, unacceptable tunnel convergence and insufficient rockbolt length—and applied the bimodal bounds method to calculate the series system reliability of a circular rock tunnel, obtaining a fairly good solution for their tunnel case. However, it is well known (see e.g. Ang & Ma, 1981) that the bimodal bounds could be wide when the probability of failure of individual LSFs are all ‘large’ (say, >0.01). In such cases, the bimodal bounds could be inappropriate.
An alternative method, based on the linearization of the LSFs, is also possible to compute the reliability of series or parallel geotechnical systems. (Jimenez- Rodriguez et al., 2006; Jimenez-Rodriguez & Sitar, 2007; Cho, 2013). However, for series systems, the idea has only been applied to very simple cases: short-term slope stability analyses of cohesive soils with two representative slip surfaces only (Cho, 2013), in which the (almost) linear nature of LSFs makes it a natural solution. In other words, the ability of the linearization approach to perform well with other common geotechnical problems is still untested. This paper aims to illustrate the ability, in terms of accuracy and efficiency, of the linearization approach to evaluate the reliability of a circular rock tunnel. To be able to compare results, the CCM used by Lü et al. (2013), will be re-applied in this paper, by considering different types of rock mass, support strategies and LSFs.
A business center’s 4-story underground garage in the center of Rijeka, Croatia is under construction. The excavation of an open pit for the underground garage commenced during Spring 2012 and was completed in December 2013. The construction site is located very close to an old shallow railway tunnel constructed in 1870. The tunnel was constructed without any rock mass reinforcement. The excavation pit for the underground garage is located only 6.0m from the side of the existing railway tunnel. The pit excavation had a significant effect on the existing tunnel construction; thus, a reinforcement of the rock mass using rock bolts and multi-layered reinforced sprayed concrete was embedded. In this paper we present a design for pit construction including numerical models and geotechnical analyses. Based on these analyses, a required reinforcement of rock mass around the tunnel and an appropriate support system for the stabilization of pit walls was designed and executed.
The construction of new buildings in old urban or built up areas is extremely demanding and, as a rule, conditioned by eliminating or minimizing its influence on surrounding buildings. The existence of empty land in old urban areas indicates that unfavorable geotechnical conditions or other construction problems exist on the site. The presence of surrounding buildings and the limitation of space for construction site organization represent the major problems during the design and construction of new buildings (Arbanas et al. 1994). The lack of necessary space requires urban planners to establish all auxiliary facilities, such as garages, underground. In this paper business center’s 4-story underground garage in the center of Rijeka, Croatia and its construction’s effect on surrounding buildings and underground facilities will be described and analyzed. The construction site is located near an old shallow railway tunnel. This railway tunnel was constructed in 1870 and is still in use. The tunnel was constructed without any rock mass reinforcement with only 5.0 to 9.0m of overlaying in poor limestone rock mass.
In this paper, a rigorous analysis by a 2D finite element code of the influence of rock mass fracturing and the influence of the water seepage in the design of a tunnel support, excavated by N.A.T.M. are introduced. Also, the existence of longitudinal joints in the rock mass is their influence in the variation of the safety factor of the primary support is taken into account. The rock mass formation is constituted by schists and shales. The field case of reference is the Prado Tunnel in the High Speed Line between Lubián and Ourense (North of the Spain). The rock characterization has been done by the Bieniawski’s rock index RMR and Barton’s index Q. Also the definition of the Hoek and Marinos GSI index is employed. The range of corrections of the RMR and Q index, originated by the influence of the analyzed factors, is determinated by the numerical simulation.
1. Introduction and Background
Two factors considered at geomechanical classifications that include direct support application criteria in tunnels are the water inflow existence and the joint disposition towards the tunnel axis alignment.
Although according to Hoek “water pressures are generally not too serious a problem in underground excavation engineering”, it must be considered to get the support in underground excavation. Equally, the orientation of joints must be considered over all when direction of discontinuities coincides with tunnel axis.
With regards the water inflow, Hoek & Marinos published in 2004, that when assigning support not only the flow net but also the joint strength reduction should be considered (GSI Index) if the joint fill or the formation are sensible to water as slate, schist, phyllite etc. Likewise the joint disposition, which is taken into account on Beniawski RMR, as well as Barton’s Q index application criteria, as reflected on 2013 update performed by N.G.I.
Uncertainties in the geological and ground models prior to construction lead to a residual risk during construction. As well the ground properties, as the exact location of the single units are not known precisely during the design. In addition, factors influencing the ground and system behavior, like ground stresses and ground water conditions can be estimated only in this stage. The consequence is a residual risk during construction. To minimize the risk, an observational approach is required. This involves sound preparation during design, and special procedures to be followed during construction. Part of these procedures is the geotechnical safety management, which will be presented in this paper.
The inherent uncertainties in the geological and ground models do not allow a precise prediction of the ground and behavior prior to construction. The associated residual risk has to be managed during construction by applying an observational approach. This includes assessing the possible range of behaviors during design, assigning appropriate construction measures to the ground behaviors under consideration of the requirements, and development of a monitoring program, which can capture the expected system behaviors. For managing the risks during construction a so called geotechnical safety management plan has to be established. The plan includes a listing of safety relevant issues for each section, predicted behavior, as well as limits of acceptable behavior in the form of warning and alarm criteria. In addition actions to be taken in case of reaching one of the warning or alarm criteria have to be specified, as well as the responsibilities clearly defined. An essential part of the safety management during construction is the implementation of an up to date monitoring program, as well as advanced evaluation and interpretation of the measurement data. An important role in the safety management plays the prediction of ground structure and expected behavior ahead of the face. This allows timely adjustment of the construction, and continuous comparison of the observed behavior with the predicted.
In this study we compared the evolution of two types of precursory indicators over time: precursory rockfalls and pre-failure deformation. Our study was based on a multi-temporal comparison of LiDAR data using a Terrestrial Laser Scanner (TLS) Ilris 3D (Optech). The analysis was performed in a pilot study area located at Puigcercos scarp (Catalonia, Spain) using 19 different fieldwork campaigns carried out during more than five years (2143 days). Our results shows that: (a) the temporal evolution of the deformation show an exponential pattern with an acceleration before the final rupture; and (b) analogous acceleration shortly before the final failure was observed analyzing the evolution of accumulated precursory rockfalls events over time.
Failure forecasting analysis has considerable benefit from the study of precursory indicators prior to the occurrence of larger failures. Some examples of precursory indicators described in literature are surface deformation (Abellán et al., 2010; Royan et al., 2013), minor scale rockfalls (Rosser et al., 2007) and micro-seismicity (Amitrano et al., 2005). Deformation is the most used indicator to predict unstable areas in rock slopes and it could be detected by several methods, including GPS (Crosta and Agliardi, 2003), extensometer (Crosta and Agliardi, 2003; Rose and Hungr, 2007) and LiDAR (Oppikofer et al. 2008; Abellán et al., 2010; Royan et al., 2013). Minor scale rockfalls leading to greater failures were back analyzed by Rosser et al. (2007) using Terrestrial LiDAR data, but their ability to forecast need to be studied in greater detail. Moreover, the analysis of different types of precursory indicators coexisting in the same area (e.g. pre-failure deformation and precursory rockfall events) has not been studied in detail yet.
We present here a comparison between temporal evolution of two precursory indicators (pre-failure deformation and precursory rockfalls) detected in the pilot study scarp of Puigcercós (Catalonia, Spain). Both indicators roughly followed the same pattern in their temporal evolution, i.e. exponential acceleration before the final rupture. Furthermore it has been noted that precursory rockfalls are detected in previous periods than pre-failure deformation.
In this paper, we present the rock mechanics tests performed at Laboratory of Lyons of the CEREMA (known at the time as CETE de Lyon), following the fire of September 2008 in che Channel tunnel. Whereas chalk marl was unaffected by the fire even in the most damaged zone, several laboratory test were performed, during the shutdown of the north tunnel.
At the request of Eurotunnel society, we determined that the rock behavior was characterized by an extreme sensitivity to water content:
• Young Modulus and Uniaxial Compressive are function of water content , with the same power law.
• Unless conserved in a fully saturated atmosphere, chalk marl of the Channel tunnel loses water at an important hourly rate and is totally dry in a few days.
This experience was useful for the SAFE project: in the first major modification to the Tunnel infrastructure since it was built, 4 SAFE fire-fighting stations were built. They became operational at the end of 2011. Those stations permit now to contain a major fire. The results of rock mechanics tests performed on samples were conform to the observation made a few months before, in a zone characterized by a higher water content.
On the 11th of September 2008 a fire broke out on a HGV(Heavy Goods Vehicle) carried by a shuttle coming from the UK side through the North tunnel. The fire lasted 20 hours. Civil works damage investigation were performed in a very short period of time on the lining (Delga, Lévy, Ducroq, & Huchon 2010). Most of the investigation concern the concrete lining, as the rock mass was unaffected by fire or its aftermath, as no lining collapse occurred even in the most damage zone.
Dynamic fracture plays a vital role in geotechnical problems. Limited attempts have been made to measure dynamic parameters of straight notch cubic sample under impact. This is due to some difficulties in preparation of samples and high accuracy needed for its testing. To solve these difficulties, it’s better to use straight notch core specimen under impact. Explanation of dynamic crack propagation by numeric analyses is limited. Among numeric methods, extended finite element method is an effective way to study dynamic fracturing. This study used the X-FEM software (ABAQUS) to create a 3D model of dynamic crack propagation of two samples; straight notch cubic and straight notch core specimen under impact load; then results obtained from ABAQUS are compared. Present study showed that Dynamic toughness for core specimen is lower than cubic specimen, Dynamic stress intensity factor for core specimen increases linearly but for cubic specimen is oscillating before fracture initiation.
Fracture mechanic has been suggested as possible tool for solving a variety of rock engineering problems, such as rock cutting, hydro fracturing, explosive fracturing and, rock stability and, based on the extension of Griffith theory and Irvin’s modification (Chen, Pan & Amadei 1998). Concept of stress intensity factor K (SIF) has been introduced by Irvin (Mohammadi 2008). Fracturing may take place under static or dynamic condition. Earlier measurements of rock fracture toughness followed the ASTM-E399 standard method. Because most rock are brittle, fatigue pre-cracking required in ASTM standard has been found to be very difficult to produce (Chen, Pan & Amadei 1998). To solve rock fracture problems, The International Society for Rock Mechanics (ISRM) recommended three suggested methods by core based specimens; for determining static fracture parameters (Iqbal & Mohanty 2007). Dynamic fracture plays a vital role in geotechnical applications frequently encountered in various engineering problems; including blasting, protective design, rock burst, projectile penetration and seismic events (Zho, Xia & Li et al. 2012, Chen, Xia & Dai et al. 2009, Dai, Xia & Zeng et al. 2011). These processes are governed by rock dynamic fracture parameters, such as fracture initiation toughness, fracture energy and fracture velocity.