The results of a testing programme involving the measurement of shearing resistance of plane and rough joints are given. It was found that the coefficient of friction measured on plane joints may be misleading, as they appear to be stress dependent. The values obtained for three rocks of different origin and strength were measured to be equal. The angle of residual shearing resistance, on the other hand, derived from testing natural or artificial joints may be more realistic. Account of a testing rig for torsion is given.
The significance of the deformation and strength properties of joints in rock masses was brought into focus in the mid sixties with the advent of computer solutions. Later referred to as stiffness in vertical and shear directions, the knowledge has came a long way from Patton's artificial joints to the current discussion of scale effects linear and non-linear have been proposed, and it is now generally accepted that the shearing resistance of a joint can be simply expressed as a function of its basic residual friction angle, joint wall compressive strength with respect to normal stresses and its surface topography. The presence of gouge and cleft water further complicates the problem. An attempt to look into the problem was made by comparing the behaviour of different rocks under identical test conditions in this investigation program. Dry and unweathered samples were tested to reduce the number of variables.
Three samples of rocks of different origin were selected from Portugal: a·A milky white holocrystalline marble from Vila Viçosa, containing rare fissures and quartz. Homogenous, appeared to loose its cohesion upon heating to 550C. b. A white, compact, calcitic limestone from Alhandra with medium size crystals. Shows strong reaction to acid. c. A giorite from Malveira da Serra containing 65.5% feldspar,9.8% pyroxene,9.3% amphibolite,3.3% biotite and 2.3% quartz.
Initially an attempt was made to measure the basic angle of friction Φb. Three different surfaces were selected for this purpose. Prisms or cores were cut and machined to a perfect fit and were sandblasted, scratched with No.80 sandpaper perpendicular to the direction of shearing, and polished with No.600 paper. Samples for triaxial cell testing were cut at inclinations Ψ= 45 and 55 degrees. The second stage of testing was implemented on prepared ‘rough’ joints. Prisms of rock with areas 60 to 100 sq. cm. were split along pre-indented lines for direct shear testing. For torsion testing NX cores were initially drilled through the centre at a diameter.
The surface topography of each sample to be tested was measured along two axes to obtain maximum profiles. A surface measuring system was developed with magnification ratios up to 200 which gave a sufficiently accurate profile, yielding first and second order roughness values.
A large number of tests to measure Φb provided data showing a wide degree of scatter. In marble, where most testing was performed scatter increased as surface changed from sandblasted to No.80 to No.600.Surface damage increased as normal stresses were increased. This was more pronounced in triaxial samples where σn went to excessive levels. Many samples were found to have failed in tension following the test. Stick-slip effect was observed in the triaxial tests.
There is now general agreement that the shearing resistance envelope of a natural joint can best be express in curvilinear form. The hypotheses put forward by Barton and Ladanyi Archambault appear to have receive t wide application. In the second par of testing, the shearing resistance of fresh joints was investigated.