PREDICTABILITY OF BARTON'S JOINT SHEAR STRENGTH CRITERION USING FIELD-IDENTIFICATION PARAMETERS

Abstract

A series of direct shear tests have been performed in an attempt at assessing the predictive capability of Barton's joint shear strength criterion derived from field-identified parameters. Ten rock types have been tested, including basalt, two marbles, three granites and four sandstones. Testing on saw-cut surface specimens determines the relationship between the basic friction angle (φb) and the rock compressive strength (UCS). Testing on specimens with tension-induced fractures yields joint shear strengths under different JRC's, for use in the verification. The results indicate that Barton's criterion using the field-identified parameters can satisfactorily predict the shear strengths of rough joints in marbles and sandstones from all source locations, and slightly over-predicts the shear strength in the basalt specimens. It cannot however describe the joint shear strengths for the granite specimens. This is probably because the saw-cut surfaces for coarse-grained and strong crystalline rocks are very smooth resulting in an unrealistically low φb. Barton's shear strength is more sensitive to φb than to UCS and JRC. For all sandstones the φb values are averaged as 33 ± 8 degrees, apparently depending on their cementing materials. The averaged φb for the tested marbles and for the limestone recorded elsewhere is 35 ± 3 degrees, and is independent of UCS. The φb values for other rock types apparently increase with UCS particularly for very strong rocks (R5 and R6). The factors governing φb for crystalline rocks are probably crystal sizes, mineral compositions, and the cutting process, and for clastic rocks are grain size and shape, and the strength of cementing materials.