Simulating progressive failure in fractured saturated rock under seepage condition using a novel coupled model and the application

Abstract: Rock mass is an inhomogeneous medium containing a large quantity of fractures and defects. The coupling interaction between seepage evolution and crack propagation is greatly significant to rock projects. In this study, a novel coupled model involving fluid pressure and mechanical damage is developed in FLAC^3D. Both the pre-peak softening and post-peak mechanical degeneration of elements are considered based on their different failure types, i.e. tension failure, shear failure, and their state combinations. Investigations are carried out on the specimen with two pre-fractures and 5% defects derived from modelling cement mortar. Four test schemes of dry and saturated specimens are conducted under uniaxial and biaxial loadings. The specimens' failure process under seepage could be divided into three stages. The failure process is basically different from that under dry conditions. Both compressive and residual strengths have declined significantly, by more than 10 and 15% respectively compared with dry specimens. The volume expansion of the specimen has remarkably increased, by more than 320%. Besides, the results are well consistent with dry experiments on cement mortar, which demonstrates feasibility of the method. Finally, the proposed model is applied to explore seepage evolution, rock mass deformation and lining stability of Jiaozhou Bay subsea tunnel within excavation. A new generation method of fracture networks is developed and the simulated results match well with field monitoring data. Article highlights: The novel coupled model considers both pre-peak damage and post-peak degeneration of rock.Principles are based on elements' failure type, i.e. tension failure, shear failure, and their combinations.Previous studies are based on elastic–plastic models, which are improved to be elastic-brittle and suit brittle rock mass better.Weakened elements are developed to simulate natural defects in real rocks.A new generation method of stochastic fracture networks is developed for engineering practice.