An underground powerhouse chamber at chibro was constructed about two decades ago in dolomitic-lime stone of lower Himalayan region. Long-term instrumentation was planned during construction to check the safety and adequacy of the support system and to monitor the post construction behaviour of the excavation. The present paper essentially ?eals with the analysis of ten year data. The observed roof support pressure has been compared with the support pressures estimated from different empirical theories. The time-dependent effect has been noticed significantly where the rock mass is saturated due to seepage problem and near thick plastic shear zone.
Construction of underground excavations for hydroelectric projects in the lower Himalayan region is a challenging task due to support problems under complex hydrogeological conditions and tectonic influences. Further, these problems could be attributed to the non-homogeneous and anisotropic nature of rockmass and their time dependent behaviour. In this context, the Chibro underground powerhouse complex has set a major precedent by being the first venture of its type in the lesser Himalaya. The 240 MW Chibro underground power station exploits the drop of about 124m along the first loop of Tons river, a tributary of the Yamuna between Ichari and Chibro, which is part-I work of Yamuna Hydroelectric Scheme Stage-II. This was the first venture of its type in the lesser Himalaya and was necessitated because the location of a surface power station would have involved large scale excavation of steep slopes. Finally, the powerhouse complex was sited in a band of limestone which has horizontal width of 193 to 217m. The complex comprises a network of excavation for the machines, transformer, turbine inlet valves and control room and also provides operating galleries and hydraulic connections to the Part-II. This latter stage involves a 120 MW Khodri power station utilising the remaining drop of 64m along the second loop. Fig.1 shows a general layout of the powerhouse complex.
We have measured ultrasonic Rayleigh and interface waves along laboratory induced planar fractures on Anstmde limestone under dry conditions. Comparing measurements of the Rayleigh wave along rough and smooth surfaces we observe that the effect of the roughness on the surface wave propagation is to induce attenuation and velocity dispersion. The attenuation increases linearly with frequency; the phase velocity goes through a maximum at 240 kHz. We put forward a physical model describing the effect of the roughness in terms of an increase in the effective length of propagation. This model predicts a frequency dependence of an order that scales with the magnitude of the roughness. We have also compared measurements of the interface wave with measurements of the Rayleigh wave, along a single rough surface, to observe the effect of stress on the interface wave propagation. Within the range from 0 to 10 MPa, the interface wave velocity and its amplitude increased with the applied stress.
Rough interfaces, in the form of hydraulically induced fractures or naturally occurring faults and fractures, play an important role in determining the productivity and final recovery of hydrocarbon reservoirs. Moreover, stress changes, as a result of the reservoir depletion, may affect considerably their hydraulic properties. To evaluate the presence and the properties (mechanical and hydraulic) of these interfaces using seismic waves, it is imperative to understand the various modes of wave propagation that can take place along them.
Non-welded interfaces have shown theoretically to support the propagation of interface waves (Pyrak-Nolte, 1987). Experiments on roughened aluminum interfaces (Pyrak-Nolte, 1992) showed that fast and slow modes of interface wave propagated along the interface. Moreover, the predominance of these modes and their velocities were strong functions of the stress applied across the interface (i.e., of the interface stiffness). It was shown that the interface wave velocity, for example, increased from the Rayleigh wave velocity, at zero stress, to the body shear wave at high stress, and that these observations were in agreement with the theoretical expectations (Pyrak-Nolte, 1992). Furthermore, numerical modeling of interface wave propagation (Myer, 1994) showed that several modes of interface wave propagate simultaneously and that these may interfere extensively with each other.
We are not aware, however, of similar experimental work conducted on sedimentary rocks. Our intention with this work is to present experimental data of interface wave propagation on rocks and to evaluate these measurements within the context of the theoretical work by Pyrak-Nolte, (1987), experimental measurements on aluminum samples (Pyrak-Nolte, 1992), and the numerical modeling of Myer (1994).
Guyer, R.A. (Los Alamos National Laboratory) | McCall, K.R. (Los Alamos National Laboratory) | Johnson, P.A. (Los Alamos National Laboratory) | Rasolofosaon, P.N.J. (lnstitut Francais du Petrole) | Zinszner, B. (lnstitut Francais du Petrole)
The frequencies of the fundamental resonances of a suite of rock samples have been measured as a function of drive amplitude. Representative results from a measurement on Fountainbleu sandstone are reported. The resonant frequency shifts downward with increased drive amplitude exhibiting a softening nonlinearity. The traditional theory of the nonlinear elastic response of rock is reviewed. When applied to resonant bar measurements this theory predicts qualitative and quantitative features that are markedly unlike experiment. The new paradigm introduced by McCall and Guyer (1994) to describe the nonlinear behavior of consolidated materials is reviewed. This paradigm is applied, using extant stress-strain data on Berea sandstone, to describe resonant bar measurements. Good qualitative and quantitative agreement with the experimental observations is found.
The traditional theory of elastic wave propagation in a nonlinear material is based on expressing the energy density as a function of the scalar invariants of the strain tensor. Landau and Lifshitz (1959) find the equation of motion for the displacement field u from
[Equation available in full paper] (1)
[Equation available in full paper] (2)
[Equation available in full paper] (3)
where å is the energy density, p0 is the constant mass density, ó is the stress tensor and ª is the strain tensor,
[Equation available in full paper] (4)
The constants ì, K, A, B, and C are found from experiment; for example, K is the bulk modulus of the material. This formulation, in which the stress is assumed to be an analytic function of the strain, has been very successful in describing the dynamics of a wide variety of materials and has been extended to describe the dynamic elasticity of inhomogeneous, consolidated materials such as rock. It is well known, however, that rocks have a stress-strain equation of state with hysteresis and discrete memory (Holcomb 1981). Thus the stress is not an analytic function of the strain. The traditional formulation does not provide a consistent theoretical framework for the description of the elastic properties of rock.
Recently McCall and Guyer (1994) introduced a new paradigm for the description of the elastic properties of rock and other consolidated materials. The central construct of this paradigm is Preisach-Mayergoyz space (P-M space) in which the response of the microscopic mechanical units in the rock, collectively responsible for the rock's macroscopic elasticity, is tracked. Given the density of mechanical units P(Pc, Po) in P-M space, where Pc and Po are pressures characterizing the hysteretic response of an individual unit, one can forward model (calculate) static and dynamic rock properties. Hysteresis with discrete memory, harmonic generation, nonlinear attenuation, and Mdynamic > Mstatic, where M is a modulus, are direct consequences of the model. Equally importantly the P-M space picture provides a recipe for experimental determination of P(Pc, Po). Thus the paradigm gives a complete description of rock elasticity; suitable laboratory measurements lead to p(Pc, Po), which, in turn, allows the prediction of all static and dynamic elastic properties.
The purpose of this paper is to demonstrate quantitative application of this paradigm. In this paper we describe a resonant bar experiment, the traditional theory of this experiment (giving an un-suitable answer), and the theory of this experiment using the new paradigm and an empirically determined p(Pc, Po) (giving results in qualitative and quantitative accord with observation).