BEHAVIOUR OF UNDERGROUND EXCAVATIONS IN SQUEEZING GROUND CONDITIONS

Abstract

Increase in population has broadened the supply-demand gap for energy. It has forced mankind to construct more number of hydro-electric power projects. On the other hand, large traffic has increased load on the roads and hence the effective travel time, especially in the hilly regions, has increased. Construction of hydro projects and shortening the routes in hilly regions involve tunnelling. In addition to this, other facilities like underground repositories for burial of high level nuclear waste (HLNW), safe storage of warplanes, missiles and explosives for defense purposes, storage of petroleum products and underground research laboratories etc. involve large underground excavations. Therefore, a silent tunnelling revolution is presently going on in India. Most of the underground excavations are carried out in the Himalayan region, which is extremely fragile and exhibits complex rock mass behaviour. High in-situ stresses, which are very common in the Himalayan region due to continuing tectonic activity, affect tunnelling in the form of time dependent tunnel deformation resulting in to large convergence of rock mass in to the tunnels. These time dependent large deformations associated with tunnelling is known as squeezing. Such deformations may terminate during construction or continue over a long time period, if adequate supports are not installed in time. Currently no methods are available which could be used with confidence in the field to assess the squeezing potential. The earlier research work relevant to the area of present study has been reviewed and discussed in detail in Chapter 2. Important outcome of some of these studies are: Wood (1972) suggested that grounds undergo squeezing, if the ratio of unconfined compressive strength of rock mass to overburden stress assumes a value less than 2. Daemen (1975) reported that deformations in tunnels are due to relaxation of broken zone and progressive reduction in residual strength of the rock has a significant influence on support pressures. It was further suggested that flexible supports are preferable in squeezing ground condition as these mobilize lesser rock pressure. Dube (1979) reported that radius of the broken zone varies in the range of 2.5 - 4 times the radius of tunnel in squeezing ground conditions. In-situ stresses are the critical parameters iv affecting the geometry of the broken zone, support pressure and the tunnel deformation. It was also suggested that Barton’s approach for predicting the support pressure needs a new parameter, viz the tunnel size, especially in squeezing conditions. Jethwa (1981) modified Daemen’s (1975) approach for predicting the support pressure after incorporating the effect of advance of tunnel face and commented that Q-system provides reliable estimates of rock pressure for tunnels in non-squeezing ground only. None of the classification systems is as such reliable in squeezing ground conditions. Verman (1993) developed empirical correlations for assessment of support pressure and tunnel deformation in squeezing ground condition involving ten parameters. Goel (1994) proposed empirical correlation after introducing Rock Mass Number, N for assessment of the squeezing ground condition. Correlations were also proposed for support pressure and tunnel closure for both non-squeezing and squeezing ground conditions. It was pointed out that the effect of tunnel size on support pressure is insignificant in case of non-squeezing ground conditions but it becomes quite significant in squeezing ground conditions. Choudhari (2007) proposed an elasto-plastic closed form solution for predicting the tunnel deformation using critical internal pressure as a parameter for circular tunnels and concluded that rock mass behaves anisotropically when in-situ stress ratio, k ≠ 1 and closed form solutions are no more applicable. Singh et al. (2007) suggested a critical strain parameter as an indicator to quantify the squeezing ground potential of the rock mass around tunnels. Lian-Chong et al. (2008) suggested that initiation of creep failure is governed by the ratio, k of the far field stresses and the creep failure initiates always in the direction of the minimum far field stress component since in that direction, the octahedral shear stress reaches the highest value. Barla et al. (2011) optimized the yield-control support system in squeezing grounds to enhance the rate of tunnel advance. Cantieni et al. (2011) suggested that if the ground exhibits a moderate time-dependent behaviour and the effect of the support measures is taken into account, the prediction of v deformation by core extrusion measurements is feasible. On the other hand, if the ground behaviour is pronouncedly time-dependent, tunnel deformation prediction becomes very difficult, because core extrusion is governed by the short-term characteristics of the ground, which may be different from the long-term properties which control the final convergence. Scussel and Chandra (2013) developed expressions based on elastic-plastic theory for the prediction of tunnel support pressure for both non-squeezing and squeezing grounds under conditions of hydrostatic in-situ stresses. Keeping in view the above discussion, following major objectives of the present study were set forth:  Study of existing tunnel case histories for understanding state-of-art of knowledge,  Collection of required data, viz., diameter and depth of tunnel, in-situ stress, rock mass quality parameters (joint properties, uniaxial compressive strength and rock mass number) and details of supports installed in tunnels from case histories,  Collection of data acquired on the basis of in-situ instrumentation and monitoring like deformations, support pressures etc.,  Development of empirical correlations for prediction of ground conditions,  Development of empirical correlations to predict tunnel deformation and support pressure for non-squeezing and squeezing ground conditions. In the present study, geo-mechanical data of 366 tunnel sections from 24 case projects located in India and other countries was collected and analyzed. Out of these, fourteen projects are located in the Himalayan region. These projects lie in India (states of Uttarakhand, Himachal Pradesh, Jammu & Kashmir, Arunanchal Pradesh and Manipur), Nepal and Bhutan. A few projects considered for study are located in the states of Madhya Pradesh, Maharashtra, Karnataka and Kerala. Four case study projects, namely from France-Italy (border), Turkey, Norway and United Kingdom were also included in the study. These data were used to develop empirical correlations for predicting the ground behaviour (squeezing, degree of squeezing, non-squeezing, self-supporting, and rock burst). Correlations have also been developed for predicting – a) tunnel deformation in squeezing and non-squeezing conditions and b) tunnel support pressure in both squeezing and non-squeezing conditions. If the behaviour of the ground is known prior to excavation vi of the tunnel-face, necessary preparations in form of excavation strategy and support installing strategy can be worked out. These correlations have been developed using joint factor (Jf), rock mass quality (Q) and rock mass number (N) separately. Other influencing parameters like tunnel radius, a; horizontal and vertical in-situ stresses; uniaxial compressive strength, σci of intact rock; and support stiffness, K have also been considered for the study. Correlations developed using Jf for prediction of tunnel deformation and support pressure have been found to fit best with the observed values in the field and have been recommended for use in the field and also in the design office. Correlations developed for prediction of ground conditions (squeezing and non-squeezing) and prediction of tunnel deformation have been validated by comparing with the respective conditions and values observed in two Himalayan highway tunnels (Escape tunnel and Main tunnel) located between Chenani and Nashri villages on national highway (NH)-1A in Jammu & Kashmir in India. Some sections of escape and main tunnels have been numerically modelled using Phase2 code and the deformation values obtained from modelling have also been compared with the predicted and observed values of tunnel deformation at respective tunnel sections. The observed values were found in good agreement with the values predicted by recommended correlations. Parametric study was also carried out for the correlations recommended for prediction of tunnel deformation and support pressure. It was observed that there is no influence of tunnel radius on the tunnel strain and the support pressure in non-squeezing ground and also on the support pressure in the squeezing ground. Further, a significant influence of uniaxial compressive strength of intact rock, σci has been observed for weaker rocks (σci < 20 MPa) at higher tunnel depth in squeezing ground condition. The correlations recommended in the present study are valid for tunnels excavated by drill and blast method. The correlations recommended for prediction of tunnel deformation in squeezing grounds, are valid for the cases where radial tunnel strain exceeds value of 1%. On the other hand, correlations recommended for prediction of tunnel deformation in nonsqueezing grounds are valid where radial tunnel deformation is below 1%.