HYDRODYNAMICS

— ESSENTIAL TO UNDERSTANDING GEOTECHNICAL BEHAVIOUR —

Finding the need . . .

A real marine construction challenge is to build dry land above the waves simply by dumping dredged sand onto the ocean floor. Experience shows that this procedure results in the sand forming a very flat (~ 3°) underwater cone; the fill is weak, and unreliable. So this method cannot be contemplated at present for important permanent structures like earthdams, causeways, or breakwaters. Hodge wanted to solve this problem. And this, his desire, lead to exposing the shortcomings in geotechnical theory - it's lack of computational utility and it's faulty theory.

The faults in current geotechnical engineering theory have their origin in the academic separation of Soil Mechanics and Fluid Mechanics into discrete disciplines of Civil Engineering. Basically, all that was necessary for Hodge to do was to reintegrate these sister-disciplines. In consequence, it became possible to move ahead with honest conviction. The following discussion of saturated soil-structure behaviour under dynamic/pulsating loading is recounted here as an example of how the application of Fluid Mechanic principles can facilitate Soil Mechanics appreciation.

Wave Flume Modeling

This video record was taken through the glass wall of the wave flume during our hydraulic model testing at the NRC laboratory in Ottawa. Here you are watching the liquefaction failure of a submerged sand embankment and its subsequent recovery of stability, albeit at a lower elevation and side slopes.

The berm was constructed by underwater sandfill placement, but here, without the application of inner drainage; consequently, the berm has flatter side slopes and higher void ratios than if it had been constructed while applying inward seepage forces.

Initially, before the waves start, the model is under simple hydrostatic pressure throughout, and that includes pore water. In other words, a standpipe piezometer placed anywhere within the berm or the surrounding water would show that the piezometric elevation (potential) was coincident with the water surface level. But once the wave train is set in motion, then the potential at any point within the system become equal to the wave water surface level directly above that point. So thereafter, the pore water pressure is made dynamic, and a function of the transient wave form.

Interpretation

These hydrodynamics are interpreted by Hodge as having the following effects on the soil-structure of the sand berm:

The leading face of the wave creates a hydraulic gradient which has two components, the vertical one acting down into the sand and the horizontal one acting towards the downstream [d/s] edge of the fill. Similarly, the trailing face of the wave creates gradients which have components, one pointing upwards and the other upstream [u/s].

Liquefaction commences as soon as the trailing face of the wave train creates a vertical hydraulic gradient in excess of the surficial sandfill's critical ("quicking") gradient. At this stage successive layers of sand are lifted off the underlying particles, thereby denying them shear resistance to motion, and making of them the solid fraction of a heavy fluid. Once reduced to a fluid, the horizontal components of gradient, both u/s and d/s, transport the liquefied sand over these respective edges of the berm.

So much for the mechanism of failure. What I find more intriguing is attempting to understand ¿ what mechanism(s) could explain the subsequent survival of that part of the berm which did not liquefy initially? And, ¿ how is it possible that the sand then in suspension can settle against a gradient which caused it to quicken in the first place? It is noteworthy that both these counterintuitive behaviours remain observable despite the continuing action of a wave train which remains undiminished in power. I believe the simplest explanation to what appears anomalous is that the newly formed soil-structure at the top of the truncated berm comes to rest at a lower void ratio than that of the original material.

I reason as follows: The hydrodynamic environment, through which the particles fall back into contact with the surface of the truncated berm, is one where a succession of reversals in the direction of the horizontal components of gradient take place as each wave's leading face is followed by its trailing face. Considering this to-and-fro movement it seems quite plausible to expect that arriving particles will seek shelter from the swaying fluid by lodgement between two secure grains in the stable soil-skeleton below. Such a need to take up a position of lateral resistance makes for tight packing density and small void spaces.

A reduction in void ratio has two effects on the hydrodynamics of the system: It increases the level of the critical gradient required to keep particles in suspension, and it greatly decreases the hydraulic conductivity (k) of the surficial layer, thereby much reducing upward flow from below. In consequence, two things happen: the downward progress of liquefaction is halted; and, the particles begin to settle out of suspension. This concept is given some support by the slope-density relationship established at NRC Ottawa, and discussed at:     Slope - Density