4.7 Laboratory-formed shear surfaces
It is axiomatic that the formation of a shear surface destroys all earlier fabric in the soil. This is a remoulding action, and it is therefore illogical to demand tests on ‘undisturbed’ specimens, preparation of which can be arduous. Remoulded specimens are adequate if a residual sứength alone is required. However, if the stress-strain behaviour as the soil passes from peak to residual is to be observed, remoulded specimens will not do. In this case only a ‘remoulded’ peak strength could be measured, and an underestimate of the full brittleness would be obtained.
A shear surface may be preformed (often termed ‘precut plane’ tests) or it may be formed by the test procedure itself. Shear boxes and triaxial apparatus can only give small strains and the preforming of a shear surface is to overcome this limitation, rather than to offer some other, positive, advantage.
Shear surfaces have been preformed by cutting a sample with cheesewứes or knives, the separated halves being polished on glass plates to simulate slickensiding. Inevitably this cannot be completely effective and even the most carefully prepared specimens do display some brittleness in shear as the plate-shaped clay mineral particles on the slip surface adopt a more complete orientation. Ideally, the test specimen would be consolidated to the desired normal stress in the shear box, and then be removed to form the slip surface. This then means that when reconsolidated prior to shearing, the shear surface is located as closely as possible to the plane of separation of the shear box halves. It will be found, however, that for practical reasons it is only worth following this sequence when testing soft and compressible materials.
Shear surfaces do form even at comparatively low strains. These have a low level of particle alignment, however, and truly residual conditions are not reached in most test procedures. Using a shear box, the two specimen halves can be racked backwards and forwards past each other until some lower bound strength is achieved. Observations of shear load V. displacement for each pass do tend to show both decreasing brittleness and decreasing strength with increasing number of reversals of strain direction, but it is not always easy to decide when a ‘residual’ has been reached. Arbitrarily defined values of total movement and/or percentage change in shear strength in consecutive reversals are often used as a criterion on which to judge the attainment of residual conditions.
With this reversal technique, there are several variants in the experimental procedure. These include measuring the load-deflection relationship on both the first pass and on the reverse direction for each stage. Alternatively, measurements of shear strength can be made in one đữection only, bringing the box halves back to the starting position by hand, or at a relatively fast rate using the drive of the machine. My personal preference is for the latter method, with time allowed for re-con$oỉidatíon after each reversal. However, reversal tests are rarely carried out in my laboratory as Ĩ prefer to use the ling shear apparatus when determining the residual strength of clays.
It is only by the use of a torsion or ring-shear device that we can achieve large enough strains to produce real residual conditions in the laboratory, starting from an initially unsheared specimen. (Bishop et al., 1971). This facility to achieve almost unlimited strain frees the user of a ring shear apparatus from a number of the constraints that beset soil testing generally. For example, since a test can be of whatever duration the user chooses, and sooner or later it must be fully drained at whatever strain rate is chosen, one of the most tricky aspects of the measurement of peak strength, or even of residual strength in a limited strain apparatus, is eliminated completely.
4.8 Experimental procedures for the ring shear test
Imposing a shear surface on a clay soil completely changes its initial fabric in the vicinity of the shear surface. There is therefore little or no merit in attempting to preserve such an initial fabric or structure in the test specimen, unless it is specifically desired to examine the process of shear surface formation. Remoulded specimens are quite adequate for residual strength determination alone.
It is usually convenient to remould soil specimens at moisture contents of the plastic limit or less. After all, the shear surface formation process is a result of soil brittleness: and this may only be manifest at moisture contents lower than the plastic limit. When wetter, the samples can extrude very easily from their container under consolidating loads. Bishop et al. (1971) have described a mechanically elaborate ring shear machine in which the gaps between the upper and lower rings may be opened and closed. This can accept fairly wet specimens. Shearing in this device takes place through the mid-height of the specimen. In contrast, the simple ring shear device devised by Bromhead (1979b) needs a drier sample to prevent eariy extrusion, and shears close to the upper loading platen. It is often found that smaller strains are required to develop the residual strength in this machine since shearing takes place through the strongly remoulded upper part of the specimen.
The following description is specifically dứected to the use of the simpler device, which is in daily use in the laboratory at Kingston Polytechnic, Figure 4.2, and in a number of academic and commercial establishments elsewhere in the United Kingdom. There are rather fewer examples abroad, but I have corresponded with users throughout much of the English-speaking world. The test has been incorporated into the 1990 version of British Standard 1377: Methods of testing soil for engineering purposes, where it is referred to as a ‘total stress’ test. The total stress and effective stress are, of course, identical in a test in which the porewater pressures are maintained at zero.
Soils may be kneaded into the sample container with the fingers or rammed into position with a wooden spatula. A short length of dowel makes a convenient rammer. Final trimming flush with the surface of the container is done with a palette knife. This has the added benefit that it begins the process of orientation of the mineral particles close to the eventual shearing surface. It is not detrimental that the beginnings of an orientated zone close to the top of the specimen are made, so that an undisturbed strength determination is ruỉed out, since all ring shear devices are totally useless for measuring the peak strength of the soil. This is because with the different strains at the inside and outside radii of the specimen, a form of progressive failure across the specimen takes place. This just about defies analysis, and all that is obtainable by way of a peak strength from this apparatus is some weighted average between a partly initially sheared remoulded strength and the residual.
Initial shearing of the specimen may be performed by setting the machine to a high rotation speed or by use of the manual control Such high rates of shear cause substantial extrusion in the simple device (as indeed it does in the Imperial College/Geonor device if the confining ring gaps are left open), and although a shear surface is usually formed by this rotation, it may not have the desired properties as a result of undrained porewater-pressure generation or because viscous interparticle drag forces have prevented the formation of an ideal, strongly panicle orientated, low strain rate, shear surface. A period of slow shearing is usually requữeđ to complete the formation of this feature coưectly.
It has been found convenient for routine tests at Kingston Polytechnic to set up a test in the late afternoon, and to observe only the early stages of this shearing, leaving the sample to shear unattended overnight. If the test rate is set to allow perhaps three revolutions to take place, then residual conditions will be approximately achieved. A number of tests have been automatically recorded throughout this shear surface formation stage, and they all show that towards the end of the allotted time, the decrease of torque through the specimen with further deformation slows down to an almost imperceptible level. One is left with the problem of ascertaining that a true residual has been reached, solutions to this are discussed below. Also, the selection of a strain rate for the remainder of the test is different to that for other soil tests: the procedure is explored in the following section.
đang được dịch, vui lòng đợi..