Modelling time dependent swelling potential of mudrock using laboratory tests and artificial neural network

Abstract

The focus of this research work is modelling time dependent swelling strain and pressure simultaneously in weak rock. Various swelling potential under different boundary conditions is considered. Free and semiconfined swelling strain and swelling pressure under oedometric condition are the most proposed tests for determination of the swelling potential of weak rock. In this paper, a modified swelling strain and pressure test is introduced using springs to provide deformable axial stiffness. The swelling strain and pressure of mudrock are modelled by artificial neural network for various swellable clay, carbonate, initial water content and different boundary stiffness, by using a combination of the proposed modified test along with the existing tests. This method correlates the results of laboratory tests with an acceptable accuracy. The proper operation of the developed model for predicting the time dependent swelling strain and pressure of mudrock under deformable and rigid axial boundaries is confirmed.

Effect of bolting on roadway support in extremely weak rock

“…Pejon and Zuquette (2002) and Moosavi et al (2006) indicated that deformation of a roadway in swelling soft rock depended on mineralogy, lithology, ground characteristics, hydrology, stress state, and weathering conditions. In addition to this, using an artificial neural network, Doostmohammadi et al (2008) investigated the swelling potential of mudstone. Lo and Hefny (1996) developed rheological models, which could be used to predict the swelling effects on a circular tunnel.…”