In this paper, the author suggests to use the yield limit as the shear failure criterion for the soft rocks of plastic-failure type. A method is proposed to determine this characteristic point, which has been adopted in several projects., The causes of scattering of the in situ shear test results are investigated, and "The method of friction-shear correction" is proposed. The practical records are presented to illustrate the advantage of reducing the scattering and artificial errors in the results. 261 sets of in situ test result on 56 engineering projects in China are analysed. The two formulas for checking the stability of the dam against sliding are suggested. However, it is not yet applied to engineering practice due to its incomformity with the current speci' ication. But preliminary trial computations show that the formula for brittle-failure type proposed by the auther is very close to that specified by the Russian standard (CH123–60); the corresponding formula for plastic-failure type is close to the friction formula without considering the adhesion, if the height of dams is within the range of 40–110 meters and corresponding factor of safty is 1.2–1.4.
The test installation and the method of load application are similar to the conventional ones, as shown in Fig. 1. In order to utilize the rock samples adequately and for the need of using "The method of friction- shear correction" described later, it is recommened first to run the single sample shear test and then to perform the single sample sliding test on the same specimen. In the single sample shear test the specimen will only be loaded to shear failure under the last increment of normal stress but failure must not be attained under every preceeding increment. Under every intermediate step of loadings the shearing stresses are increased continuously, until the shear failure of the specimen is approached indicated by the abrupt increase of the shearing strains; then the loading is release to zero. Subsequently the specimen is reloaded with successive normal stress increments as shown in Fig. 2. The last increment of normal stresses for various specimens must not be equal but arranged in progressively increasing or decreasing way. In this way a set of ultimate shear strength results may be obtained. For each specimen; after completion of single sample shear test, the single sample sliding test is carried out. This kind of method is advantageous for utilizing the specimen cconomically, especially in cases, when the test site may only be arranged in a narrow zone. In this manner for example, a set of 5 specimens may result in one set of ultimate shear strengths(peak value).5 set of single sample shear strength (yield value) and 5 set of single sample sliding strength, accounting for 11 sets of test results. This will greatly facilitate the comparison of various results, and in particular, this in necessary for "The method of friction-shear correction" to be shown later.
(Figure in full paper)
(Figure in full paper)