The creep behaviour of a soft rock can be examined by experimental studies involving triaxial creep tests and tunnel physical model tests. The physical model tests show a close resemblance to the developed numerical analysis using fitted creep parameters determined by triaxial creep tests. At lower boundary stresses there was no obvious creep but as the boundary stresses increased the model displayed primary and secondary creep. The non-linear creep model using stress-dependent creep parameters, assigned according to the changing stress conditions in the creep process, appears to be a more rational approach to time-dependent behaviour than the traditional linear creep analysis.
Creep is a peculiar form of rock deformation that can occur in very soft or heavily fractured rocks under significant in situ stresses. In Taiwan, due to the inferior geological setting and young sedimentary rock formation, creep in tunnels is a very cormnon problem which has caused severe cost and time overruns for highway and rail tunnels which are a crucial part of the infrastructure of the country.
A research progranune was undertaken to investigate this phenomenon and one of the objectives was to develop a useful analytical procedure for designing tunnels in rocks that are subject to creep under construction stresses. The work was supported by Sinotech Engineering Consultants, Taiwan and carded out by Geotechnical Research Center of Sinotech in collaboration with the University of Bradford, UK. This research programme used laboratory testing, physical modelling and in situ measurements at runnel sites to investigate the various creep model and to validate a numerical code developed for the analysis of runnels in soft and highly fractured rocks
The paper will concentrate on the physical modelling. The testing apparatus was a biaxial system with a capacity of 2.5 MPa boundary pressure in a plane stress condition. Of particular significance were the test materials which were developed to simulate soft rocks with creep tendendes and the loading system which was an air over oil actuator.
The tunnel physical model tests show a close resemblance to the developed numerical analysis using fitted creep parameters determined by triaxial creep tests. At lower boundary stresses there was no obvious creep but as the boundary stresses increased the model displayed primary and secondary creep. The non-linear creep model using stress-dependent creep parameters, assigned according to the changing stress conditions in the creep process, appears to be a more rational approach to time- dependent behaviour than the traditional linear creep analysis.
