Modelling the shear behaviour of clean rock discontinuities using artificial neural networks
RIS ID
113097
Abstract
Since the mechanical behaviour of rock masses is influenced by the shear behaviour of their discontinuities, analytical models are being developed to describe the shear behaviour of rock discontinuities. The aim of this paper is to present a model to predict the shear behaviour of clean rock discontinuities developed by using artificial neural networks (ANN), as an alternative to the existing analytical models which sometimes require certain parameters obtained from large-scale laboratory tests which are not always available. Results from direct shear tests on different boundary conditions and types of discontinuities have been used to develop this ANN model, whose input parameters contain the boundary normal stiffness, the initial normal stress, the joint roughness coefficient, the compressive strength of the intact rock, the basic friction angle and the horizontal displacement of a joint. This proposed ANN model fits the experimental data better than some existing analytical models, and it can satisfactorily describe how governing parameters influence the shear behaviour of clean rock discontinuities. This paper also presents a practical application where the proposed ANN model is used to analyse the stability of a rock slope.
Publication Details
Dantas Neto, S. Adonias., Indraratna, B., Oliveira, D. Americo Fortuna. & de Assis, A. Pacheco. (2017). Modelling the shear behaviour of clean rock discontinuities using artificial neural networks. Rock Mechanics and Rock Engineering, 50 (7), 1817-1831.