STRESS AND DEFORMATION ANALYSIS OF UNDERGROUND POWERHOUSE CAVERN USING 2-D FEM: A CASE STUDY

Abstract

Design and analysis of underground structure such as cavern has always been regarded as one of the most difficult job for engineer. The main reason is that such problems as determination of the surrounding rock load, the force distribution and the deformation behaviour of the rock mass around cavern have not been thoroughly understood. Nevertheless, the engineers have been striving to get some empirical formulas or approximate results by way of observation, statistical analysis and numerical methods. In recent years, as the development in computer technology is advancing, numerical methods are increasingly becoming a popular engineering tool for design & analysis of underground structures and out of the available numerical methods, Finite Element Method has emerged as the most popular due to its flexibility in handling material inhomogeneity and anisotropy, complex boundary conditions and dynamic problems, together with moderate efficiency in dealing with complex constitutive models. Review of literature reveals that a number of analytical studies have been carried out to investigate the behaviour (stress and deformation) of rock medium around cavern is identified. However, most analysis of underground cavern is elastic or employs Mohr-Coulomb or Hoek-Brown criteria for elastoplastic. Hence the present study aims at providing a better understanding of the behaviour of rock mass around cavern using finite element code ANSYS in elastic and elasto-plastic with Drucker-Prager criteria approach and also to address some issues regarding the capabilities of ANSYS in modeling rock mass. In the present study, two types of rock mass domain (in elastic and elastoplastic material approach) from 2-D nonsymmetrical vertical cross section of powerhouse cavern of Cirata (Indonesia) have been analyzed The cavern is 35 m wide, 49.6 in high and 280 m long. It is located in dominantly laharic breccia. The rock mass properties from previous study have been used in the analysis. The rock support system i.e., shotcrete and rock bolts/prestressed anchor has also been modeled lime time-dependent behaviour., anisotropy of the material • and the joints or other discontinuities those materially influence the behaviour of rock mass have not been ii considered to simplify the problem. The same cavern previously studied by Reik (1986) & Kamemura (1986) using finite element analysis and LAPI 1773 (1995) employ both finite element and distinct element through UDEC code. The results of the present study, in term of the deformation around periphery of cavern, then compare with the previous study and insitu measurement's data. The results ofpresent study reveals that elastoplastic model, both of Model I & Model II, always give magnitude of horizontal stress lesser than elastic model, but not always in term of vertical stress. All maximum and minimum, horizontal and vertical stress always occurs on lower part of cavern. The analysis recognize that minimum horizontal stress is very high which seem to be unreasonable (- 6.6 Mpa tensile stress), and probably exceeds the tensile strength of rock mass. In term of rock deformation, the maximum horizontal deformation occur at lower part of cavern e.g. wall of the bench, whereas the maximum vertical deformation occur as expected, at roof of cavern. From comparison with previous study generally there is similar evolution of deformation on nodal investigated with slightly anomalies, and elasto-plastic model in present study show the results comparing well with record measurement. Overall, from deformation field Model I gives results closer to the previous study and instrumented data. On the basis of the above observation, it may be concluded that Model I and elasto-plastic material approach are better for represent the problem domain. Such a study is found to be very helpful for evaluating stress and deformation behaviour of the rock surrounding underground cavern in elastic model as well as elastoplastic model. Regarding the capabilities of .ANSYS, there is main difficulty, which arise when user wants to simulate the horizontal ground pressure for pre-existing stress state since no option in the code for that purpose. However, with the techniques such like applying the horizontal ground pressure at far-field boundary or at periphery of opening, the simulation can be done although not so representing the actual stress state of rock mass.