Granular material is better to work with than clay. It is hard to really do much wth out some laboratory testing, yet that rarely happens. Regarding cohesion: Cut slopes: Cohesion is at its maximum prior to drainage. As the soil drains, the cohesion decreases and the friction (phi) increases. For normally consolidated soils, the strenth gain from friction is greater than the cohsion loss, so the soil increases in strength. For over consolidated soils, that may not be true. The gain in friction, may be less than the loss in cohesion. This has caused slopes that were stable for long periods to fail eventually.
Fills: Clays can be used for fill, but it is harder to get good compaction. Need a "sheap foot" roller and need to pay close attenton to the water content.
Can't say athat I agee with DRC1. In general as clay loses moisture due to various causes, it shrinks to a point that cracks develop, but its cohesion, in what is left, increases.
Of the clay slopes that I have seen fail, they usually are failing due to excessive water, usually ground water from up-slope. At that point cohesion has dropped as moisture content increased.
Of course there are other factors such as pore water increasing in pressure, heavier soils, etc.
A simple rule to follow is "keep that water out" and your initial soil properties probably prevail.
On compacton of clay soil, if you compact to a density greater than the initial undisturbed density in the borrow area, eventually that soil will return to that condition. If highly compacted, it likely will swell to do so. I learned the hard way. Kind of embarasing.
Only just noticed your question of a few days ago, hope you still need an answer.
The stability of a slope or loads on a retaining wall are generally governed by the long term soil parameters. The short and long term strengths of a free draining granular soil are substantially the same. However, for a cohesive soil the characteristics change with time from a cohesive material with little internal friction to one of low cohesion and a high angle of friction.
I have been caught out in the past by doing sophisticated effective stress tests, measuring pwp over several weeks to determine the effective cohesion and angle of internal friction. Generally C' is low, perhaps 2 to 10KPa while phi' can vary. However, having carried this out on a project that went to litigation, it was suggested that it is generally prudent to assume C' is zero. This may be a worst case sceanario but it's safe.
When it comes to the analysis, the ground water conditions are probably the most critical parameter. In practice excessive ground water pressure will cause a failure regardless of the soil parameters you assume or actually exist therefore during construction these must be controlled.
Something to keep in mind - the use of c' and phi' is a mathematical representation of the soil behaviour. phi' really isn't all that straight - especially at higher stresses. At low stresses, true effective cohesion is basically zero - but the friction angle is a bit higher than in an average sense. This has been explained in several texts. The undrained "cohesion" is basically a a convenience since pore pressures are difficult to predict - and unsaturated soils makes it even more so.
It is sort of a loaded question, but I don't have much of an answer other than to say the soil physicists with their theories involve many factors, such as electrical forces within and between molecules. From what I understand it can get very involved. Thus, "cohesion" can involve several differing forces, it you will, and so describing exactly with is going on and why can be a big subject in itself. I suspect we engineers are just scratching the surface.
Going on one step, take a certain clay and place it in a potter's oven. Now what is the "glue" factor that holds that pottery together? Is it still cohesion as we would use it?
To add to the discussion,
When dealing with soft rock, such as a shale or mudstone, Does the measured 'cohesion' essentially define the extra strength of this material, considering the phi angle is much the same for 'intact' samples and remolded samples? This cohesion, ranging from quite high to lower values is present in dry to moist to saturated rocks.
As the soft rock weathers, the measured cohesion decreases, finally approaching the remolded values.
As oldestguy discusses, what is this property, which we use in the field and are so afraid (and rightly so) to incorporate into our computations. It is related to moisture content but is also related to geologic weathering. It is legitimate for us to use it (or some of it), if we have properly defined the present and future conditions, as applied to our project.
But how often can we safely predict that clay will not get wet?
I have used some cohesion for design, but usually reduce it for saturation and/or residual strength values.
Keep in mind that may people when they decide to neglect cohesion force the envelope through 0. This has the effect of increasing the phi angle of the soil. When looking at the shear values vs. the normal loads, this will give you more conservative at low normal loads, but less conservative at high normal loads. I prefer to check that the strength used in design is below the envelope for the normal load used, not just get the number of degrees and run with it.
I think the real problem is that many people do not use actual results on a day-to-day basis, and do not have a good understanding of what the data means. Then have to plug it into a program , and not know how to check the information .
Hope I didn’t derail the conversation, the opinions on cohesion interest me.
Is an intersting point OldestGuy. Of course, I've had cohesionless silty sand adhere to my boot too! But, not gravel. I think that adherence can happen at low confining pressures but it doesn't take much pressure to clean off the boot. So, while one may argue that there may be some 'cohesion' at very low confining pressures, I don't think that these low pressures really come into our analysis when we are dealing with "metres" of soil. It is a similar argument to threshhold seepage gradients.
