An investigation and review of oedometer and triaxial tests on Winnipeg clays

Thumbnail Image
Trainor, Peter Gordon Samuel.
Journal Title
Journal ISSN
Volume Title
An extensive laboratory program of small diameter CIU triaxial tests and oedometer tests on undisturbed Winnipeg clay has been executed. Blocks of clay from depths of 5.5 m, 9.0 m and 11.5 m depth were chosen for the study to substantiate and augment previous research. Significant new insights have been gained into the behaviour of the clay, and explanations have been offered for previous apparent anomalies. The traditional semi-logarithmic plot of the compression versus the logarithm of the effective stress in an oedometer test is inappropriate for Winnipeg clay. Arithmetic scale plots of compression versus effective stress are much more useful. The later show that in the recompression range the relationship is linear, but the compressibility is unusually high. After recompression a yield occurs at the preconsolidation pressure, P'c, but the higher P'c, the lesser the amount of yield. The value of P'c decreases with depth so that at 5.5 m the yield is barely noticable, whereas at 11.5 m the yield is very sharp. To increase the definition of the yield and also to complement the arithmetic scale plotting method, many of the oedometer samples were loaded with small equal increments. The more usual method is to load samples using a constant increment ratio, which suits logarithmic plotting. In triaxial compression the Winnipeg clay behaves as a lightly to moderately overconsolidated deposit, with Af values ranging from 0.0 to 1.1 The effective-stress strength envelopes are, however, unusual in shape. At low effective pressures, the envelopes are curved and appear to have zero cohesion intercept. It is demonstrated that this curvature at low pressures is likely caused by a general softening of the microstructure, rather than by failure along fissures. At higher pressures there is apparently no abrupt transition between normally consolidated and overconsolidated envelopes, even though normally consolidated and overconsolidated failure modes were distinctly different. A possible explanation for both the curvature and the lack of a transition is a high content of the swelling clay mineral calcium montmorillonite or of a similar mineral. Both normally consolidated and overconsolidated strength envelopes for this mineral are curved, and the behaviour of the natural Winnipeg clay may be similar. Although the author's envelopes are unusual in shape, the test results are highly compatible with the results of previous research. The unusual shape is in fact an explanation for the widely conflicting c' and o' parameters previously quoted for normally consolidated and overconsolidated envelopes.