Multiple-step Loading Triaxial Compression Test (ML-TCT), allows evaluating shear strength parameters, i.e. cohesion and friction angle from a single specimen. The test consists of a series of consolidation and shearing steps, and in each step the axial loading is stopped or reversed before a complete failure. The ISRM suggested method have suggested to increase axial load until the axial stress-strain curve show a horizontal tangent (Etan = 0). However, due to large plastic deformations in the specimen immediately before arriving to this critical point, the shear strengths obtained by this method are prone to be underestimated comparing with those for intact specimens. In this study, a new multiple-step loading method is proposed for the determination of the triaxial compression strength envelope. Series of Single-step Loading Triaxial Compression Tests (SL-TCTs), and ML-TCTs in increasing and decreasing confining pressure steps were carried out using a sedimentary soft rock, mudstone. Careful examination of the axial stress-strain relationship lead to the conclusion that, the point at which the secant Young's modulus became the maximum value (Esec = max) appeared to be a suitable criterion for estimating the imminent failure of the rock specimen. The failure envelopes derived using this method closely approximated that derived from SL-TCTs on intact specimens. Whereas, the shear strengths determined based on the ISRM suggested method were not in good agreement with those for intact specimens.
Multiple-step loading triaxial compression test (ML-TCT), which was proposed by Kovari in 1975 , permits determination of a failure envelope by testing a single rock specimen in a series of consolidation and shearing stages. In appropriate rocks, this technique allows the determination of strengths parameters from many fewer specimens than does conventional triaxial testing, which requires three or more specimens to determine a failure envelope. However, results of previous research works reveled that the differences in the peak strength between the ML-TCTs and the Singlestep Loading Triaxial Compression Tests (SLTCTs) can be variable. This can be explained by the fact that the operator cannot confidently recognize the peak strength and halts the test prematurely. The results depend critically on the ability of the experimenter to recognize imminent failure of the test specimens and can therefore be inconsistent between laboratories. In an effort to optimize the multiple-step loading procedure, various criteria for determining the load reversal point have been evaluated. The point most widely used by researchers is that of "a sample exhibiting failure". Kovari and Tisa (1975)  showed that the peak strength obtained in multiple-step loading tests using this point is negligibly different from that obtained by single-step tests. These results justified the use of this criterion in a number of studies. However, there is no doubt that this point is not easily identifiable during the test. Additionally, large amount of plastic deformations take place in the samples experiencing these stress magnitude. Moreover, Kim and Ko (1979)  described the dependency of the effectiveness of this method on the type of stress-strain curve.