Oxford University
Civil Engineering
Department of Engineering Science
Flocculation and consolidation of fine-grained cohesive sediment
Objectives
The objectives of this work are to obtain data on the effect of floc size and structure on the development of bed strength and on the formation of bed structure. Experiments will correlate the input conditions and floc structure to the characteristics of the developing bed.
Background
An understanding of the behaviour of newly sedimented soil is important because of its direct applications in areas such as siltation, dredging, storage of waste slurries and land reclamation. Previous work at Oxford has shown that the deposition history can have significant effects on soil structure at very low effective stress levels.
The Experimental Programme
Settling column experiments are under way at the University of Oxford. Bed properties such as density, shear stiffness, shear strength and fluidisation resistance will be correlated with floc size and deposition rate. Variation in the floc sizes will be induced by turbulence grids or stirrers installed in the sediment columns, by the choice of different input rates of sediment, and by changing water properties (pH, salinity etc.).
In order to understand the effect of floc characteristics on sediment beds it is necessary to simulate the sediment cycle, from floc creation through to consolidation, with minimal disturbance. To represent natural estuarine and coastal conditions the mud must be introduced at very low concentrations and flocs must have appropriate time and conditions to develop at these concentrations. The sediment introduction system (SIS) developed for the present experiments consists of computer controlled valves, pumps and stirrers. The SIS is illustrated in the figure above.
A known mass and density of clay is stored and kept stirred in a large mixing basin. This clay is very slowly and accurately pumped into a smaller basin, using a peristaltic pump. At the same time water from a reservoir is allowed through a pinch valve for a specified amount of time and mixed with the mud in the mixing basin. This small basin is kept stirred using a magnetic stirrer, at a specified speed. The mud is released from this basin at a very slow, but constant, rate through narrow tubes into the sediment columns. Meanwhile the water reservoir automatically refills itself from a larger source. The computer software is set up so that the system input and output are roughly equal, however there are liquid level sensors, which pause the program when the mixing basin is full, and several overflow outlets which allow water or slurry to escape in the event of blockage. In this system a known quantity of mud can initially be placed in the large mixing basin, and it can be steady diluted over several days to weeks, rather than having to refill the mud source constantly.
Floc Measurements
Floc size and settling velocity are monitored with a computerized video capture system and image analysis software. The capturing of images directly to the computer gives several advantages such as simplicity of use, more efficient computerised analysis and higher quality images. The flocs are illuminated by a laser beam sheet lying on the plane of focus of the video camera, to avoid the imaging of out of focus flocs.
sample
movie(10 Mbytes)
Density
is indirectly measured using a non-destructive X-ray technique developed
and used in many studies at Oxford. The arrangement is shown here schematically.
A highly collimated beam of X-rays is directed through the column to
a sodium iodide crystal and photomultiplier assembly, producing a count
rate which can be related to the density.
Density calibration from the count rate is made by using samples of known density. These samples are enclosed in sections of perspex column taken from the main sedimentation column and use the same soil used in the experiments. The X-ray and counting assembly are traversed vertically allowing measurements to be made along the entire column height with a vertical resolution better than 1 mm.
Shear
wave velocity, together with soil bulk density, is used to calculate
the small strain shear stiffness of the soil, which gives an indication
of the development of the soil structure. This is important in distinguishing
an open card-house structure from a more dense soil. A schematic of the
bender element system developed for this research is shown here.
Bender element transducers comprise of two ceramic plates. When a voltage is applied to the element one plate elongates and the other contracts, resulting in a bending displacement. With the bender element clamped at one end the free end vibrates. A shear wave is generated and propagates through the soil. A receiver element converts this mechanial energy back into electrical energy.
The
signals are transmitted and received by a custom designed instrument,
which consists of three main components, a function generator, a filter
and an amplifier, coupled to a personal computer. Software written in
Visual Basic keeps track of the signal received. The time difference
between the trigger and received signal gives the transmission time through
the soil, and thus the shear wave velocity.
Pore water pressures are measured using a technique and apparatus originally developed at Oxford. A single transducer is connected in turn to ports at different heights on the column wall through a central pressure multiplexer as shown above. A plastic Vyon filter is placed in each port, allowing the transmission of water pressure without allowing the sedimenting soil to escape. Each of the pressure line inputs can be connected to the transducer in turn by rotating the top disc of the CPU (see figure above) relative to the bottom disc.
One of the pressure lines is connected to a calibration tank. Starting at a base (relative 0) position this tank is raised to a height of 1 metre, at 10 cm intervals. An equation (straight line) and regression coefficients are calculated for the voltage vs. the height of the calibration tank. A calibration may be made of the pressures observed in the column to those in the calibration tank by assuming that the slope of the voltage vs. height line will be the same in both circumstances. Accuracies are of the order of 0.01 kPa, or 1 mm of head.
An up to date summary of the results is given in the COSINUS pages.
