ENCAPSULATION: A NEW CONCEPT FOR THE DISPOSAL OF CONTAMINATED SEDIMENT.

A FEASIBILITY STUDY

Ir. S.W. Davids

Prof. Ir. J. de Koning

Dr. ir. S.A. Miedema

Ir. W.F. Rosenbrand

XIIIth World Dredging Congress

Dredging for Development

7 10 April 1992

Bombay

India

ENCAPSULATION: A NEW CONCEPT FOR THE DISPOSAL OF CONTAMINATED SEDIMENT.

A FEASIBILITY STUDY.

Ir. S.W. Davids ¹

Prof. Ir. J. de Koning ²

Dr. ir. S.A. Miedema ³

Ir. W.F. Rosenbrand ⁴

ABSTRACT.

Waterways are subject to the continuous deposition of sediments and therefore need maintenance regularly. To ensure the desired water depth it is necessary to dredge the waterways. In some cases the dredged material is contaminated to such an extent that it should be isolated. Aquatic sediments may be so severely contaminated that remedial dredging is necessary.

There are many treatment methods developed to prevent the spread of contaminants into the environment. These methods vary from disposal in a contained site to total cleaning of the dredged material.

The paper describes the rheological problems to the method of inserting contaminated silt in a layer of clean silt as suggested by de Koning and Miedema in 1989/91 [9,10]. The advantage of this method is the use of the isolating properties of the clean silt which is surrounding the contaminated material. The property of fixing the free toxides in the consolidation water is the most important. A problem in realizing this method is the insertion of the contaminated material in the layer of silt which is assumed to be clean. For this reason a study of the rheological properties of silt was carried out. Silt is a material which can be described with the Bingham fluid model (viscoplastic). With this knowledge model experiments were carried out with the aim to verify the feasibility of the developed insertion method. The model experiments resulted in a deeper understanding of the rheological processes occurring during the insertion and gave the authors the opinion that the developed insertion method has good prospects.

¹ Student at the time of the research, Chair of the Technology of Soil Movement, Faculty

of the Mechanical and Marine Engineering, Delft University of Technology, The

Netherlands. Currently in the employ of IHC Holland n.v..

² Professor, Chair of the Technology of Soil Movement, Faculty of Mechanical and

Marine Engineering, Delft University of Technology, The Netherlands.

³ Senior lecturer, Chair of the Technology of Soil Movement, Faculty of Mechanical and

Marine Engineering, Delft University of Technology, The Netherlands.

⁴ Senior Development Engineer, Royal Boskalis Westminister n.v.

INTRODUCTION.

During the last decennium the processing of contaminated dredged materials has been the subject for many studies. At the Delft University of Technology students of the Faculties of Civil and Mechanical engineering carried out inventory studies on the subjects of the market, classification of the contaminated sediments, regulations, dredge and processing techniques and depositing methods.

One of these students, Makkink 1989 [7], carried out an inventory study concerning the classification of contaminated sediments and the available depositing methods.

This study resulted in the opinion that the method of encapsulating contaminated silt in a layer of clean silt as suggested by de Koning and Miedema 1989/91 [9,10] in the assignment of the thesis subjects for the students Makkink 1989 [7] and Davids 1991 [2,3], is a promising alternative. In Holland there are some examples of existing deep sandpits that could be used for the realization of this method.

In practice this method could possibly be realized by inserting the contaminated silt as spherical bodies in a layer of clean silt.

The first objective of this investigation was to carry out a feasibility study with regard to the stability of the inserted silt bodies in a clean silt layer.

Secondly, if the method is proved feasible, a sensitivity study, which was designed to find out the influence on the shape of the inserted silt body of:

- the ratio of the strength of the inserted material and layer material

- the inserted flow

- the type and shape of the insertion mouth

The influence of density stratification, which appears in a natural layer of silt, is not investigated. However this phenomena can most probably be neglected when this method is put into practice.

CONTAMINATED SEDIMENTS.

In this context contaminants are considered to be all toxic materials appearing in the sediment in a concentration exceeding the natural concentration. In the last decennia the knowledge with respect to the effects of contaminated silt on the environment has increased.

The most hazardous contaminants are PCBs (PCB = polychloorbifenyls) and pesticides. PCBs are chemically stable organic substances, used for the instance as components of lubricants. Contaminants can be differentiated into organic and anorganic substances. Chemical substances containing heavy metals fall into the last category.

The spread of pesticides often results from the dissolving of the pesticides in rainwater which then enters the surface run off or groundwater systems. The rate of decay of the pesticides in the soil, the properties of the formed residues and the ratio between the amount of bonded and free substances are important with respect to contamination.

Contaminants which spread in the surfacewater will be transported by rivers and channels and through lakes to the sea. In the same way material eroded in the upper course of rivers will be transported downstream and settle as silt in the lower reaches.

Many contaminants are absorbed by the fine fractions of silt, thus silt has the capacity to immobilize and render harmless many toxic elements. This is a valuable property of the fine fraction of the sedimented material.

Adsorption is a reversible process which is influenced by aerobic or anaerobic environment, acidity and redoxpotential. As long as the chemical equilibrium is not disturbed the contaminants will remain bonded to the soil. The exposure of such dredged material to air disturbs the chemical equilibrium so the bondings of the toxic elements are broken (due to oxidation). This results in free movement of the toxides in the environment, making dispersion possible. This implies that it is necessary to find ways in which the contaminated material can be placed in an anaerobic environment.

Consolidation of the settled silt in the storage depot causes a higher pore pressure than the waterpressure outside the depot. This pressure gradient results in pore water flow causing dispersion of contaminants. In a depot above the waterlevel this pressure gradient is increased by the hydrostatic pressure difference in waterlevel between inside and outside. So the decision to use underwater disposal sites seems to be justified. Calculations indicate that the actual amount of consolidation water which passes through the depot boundaries is so small that an additional clay layer does not contribute to the isolation. In other words, contaminated silt isolates itself.

Adsorbed substances can only be transported by means of the erosion or by uptake by biological life. Makkink 1989 [7] stated that these means of transportation are not applicable in the case of a contaminated silt body for example spherical shaped) inside a clean layer of silt. From this the conclusion can be drawn that only the transport of dissolved substances in the porewater has to be considered. For the adsorption ratio of the following equation can be used:

(1)

with: C_a = the concentration contaminants adsorped in the dry soil [g/g d.s.].

C_o = the concentration contaminants dissolved in the pore water [g/ml pore water].

K_d = distribution coefficient solid-liquid [ml water/g d.s.].

RHEOLOGY.

Silt is a homogeneous mixture of water and very fine material. This mixture contains mineral particles smaller than 0.06 mm. The distribution of solid particles in the liquid is equable. This is largely caused by the Brownian movement of the particles of colloidal size, which is in order of magnitude less than a micron, or by a weak turbulence (small vortices) of the veryfine particles with a very small settling velocity.

These mixtures behave like a liquid with a specific density equal to the sum of specific density times concentration of liquid and solid particles. They have viscoplastic properties with a yield strength, meaning that the material behaves like a solid until the shear stress exceeds the yield shear strength. As soon as the shear stress has exceeded the yield shear strength the material behaves like a fluid.

From that moment on the suspension will flow. At low Reynolds numbers laminar flow will occur. Above a certain critical Reynolds number the flow will become turbulent.

Silt can be represented by a fluid model named: Viscoplastic. The theoretical background has been studied by Bingham so fluids behaving according to this fluid model often are called Bingham fluid. Figure 1a gives a graphical representation of the behaviour of Newtonian and Bingham fluids in a shear stress/velocity gradient diagram. Figure 1b shows the resulting apparent viscosities for both types of fluids.

Mathematical the viscoplastic fluid can be described as:

(2)

with:

and:

t₀ represents a finite yield stress which must be exceeded to initiate flow.

In real materials, the actual yield stress is a rather difficult quantity to discern from laboratory data. It should represent the value of t as dv/dy approaches 0. However, for very slow flow rates the response of the shear stress changes rapidly and is usually affected by the shear history of the test sample, including placing the sample in the test instrument. Testing with gradually increasing shear strain usually fails to disclose a yield stress which is unambiguous. The technical importance of the yield stress is therefore likely to be more apparent then real. This might be caused by the solid properties of the substance which are not present in the Bingham fluid model. However, the substances of interest do not return to a plane surface if the free surface of the substance is disturbed.

h_B represents the so-called Bingham viscosity. Also the apparent viscosity is defined. In figure 1a a line is drawn from the origin to point A. This line through point A can be represented by:

(3)

The shear stress found according to this equation should be equal to the shear stress found with the Bingham equation. This gives:

(4)

So for the apparent viscosity:

can be derived.

(5)

Figure 1b shows the constant Newtonian viscosity h_N, the Bingham viscosity h_B and the apparent viscosity h_a. For very high velocity gradients the apparent viscosity will approach the Bingham viscosity as depicted in this figure.

ENCAPSULATION.

By using the Bingham properties of silt in a positive manner, contaminated silt can be encapsulated in a layer of clean silt. Because of the presence of a yield stress in Bingham fluids the contaminated silt body will remain intact as if it were a solid body.

The basic idea of placing contaminated silt in a layer of clean silt can be put in practice by using a dredging device which has been adapted for this job. The vessel must have an insertion pipe which can be positioned at a certain depth in the clean silt layer and the density of the contaminated silt. The contaminated silt can be pumped through the pipe into the clean silt layer. An impression of this method is given in figure 2.

Sites where this method can be put into practice are old sandpits where after a period of time a consolidated silt layer has been formed. Another possibility is to dredge a pit and fill it (partly) with a layer of clean silt.

As can be seen in this figure the insertion mouth can be positioned at any depth in the clean silt layer, which is important for the insertion of the correct density, but this also prevents the contaminated silt from coming into contact with the air and with the clean water. So the contaminated silt is isolated from the environment immediately during and after the insertion.

A small suction dredge with an insertion mouth placed at the position of the suction mouth, could be used. This depends, however, on the way the contaminated silt is delivered to the dredge.

ENERGY EQUATION.

In the introduction of this paper it is stated that the contaminated silt should be inserted in a clean silt layer as spherical bodies. During the insertion of material in the clean layer, the energy required is supplied by a pump. From the energy equation it can be derived that the most probable shape of the contaminated silt bodies is spherical.

From the energy preservation criteria an equation can be derived, which is valid for stationary flow with friction and external supply of heat and mechanical energy. This means that the energy balance is considered at a certain time interval Dt when material is inserted with a constant velocity.

This equation can be written as:

(6)

where: DU = internal energy

l = efficiency coefficient for pipeline resistance

DQ_w = heat flow through the system boundaries

DW_pump= pump energy supplied to the system

DE_kin = kinetic energy

DE_pot = potential energy

Equation (6) applied on the defined system gives energy flows according to figure 3.

The increase of internal energy is due to dissipation of energy by internal friction which is caused by plastic deformation and shear of the inserted material and the layer material. Because the material has a yield strength, shearing is very energy consuming. In nature many processes take place on the basis of the minimum deformation energy principle. For the material injected this implicates the occurance of plastic deformation. This natural property leads to the conclusion that the most probable shape of the contaminated silt body is globular. The minimal use of energy in a thermodynamic process occurs when the process is reversible (isentropic).

If the process of insertion is considered to the reversible it is possible to make a thermodynamic derivation of equation (6) which leads to:

(7)

This equation gives a relation between the increase of temperature, the change of potential energy, the change of kinetic energy and the increase of pressure in the insertion mouth at the end of the pipe.

To validate this hypothesis, with respect to the shape of the inserted silt body and the prediction of the inlet pressure, it is necessary to carry out a series of physical model experiments.

MODEL EXPERIMENTS INTRODUCTION.

The goal of the experiments is to make visible the development of the shape of the inserted silt body. Because silt is opaque, a transparent substance had to be found with the same model behaviour as silt. Several substitutes were investigated with respect to their fluid properties. For the model experiments hair gel produced by the cosmetic industry was chosen, because the transparency is satisfying under the conditions of the model experiments and its fluid properties match the Bingham fluid model.

The presence of a yield strength can be recognized because the substance does not return to a plane surface if the free surface is disturbed and air bubbles which are caught in the substance do not rise to the surface.

DIMENSIONLESS NUMBERS.

To set up physical experiments it is necessary to scale down the prototype dimensions to dimensions of a test stand on laboratory scale. It appears there is no theoretical calculation method to solve this problem of scaling down.

If there is no solid theoretical background, dimensionless numbers are often used to determine the model rules. Considering the parameters which are mentioned in the introduction of this paper there are two dimensionless numbers, which are of importance for the insertion process. The first number is the Reynolds number:

(8)

This can be calculated in two positions. It can be calculated in the insertion mouth just before the inserted material enters the clean layer of silt and just after the material has left the insertion mouth.

The second number is the Hedstrm number:

(9)

This number is developed for viscoplastic materials because it contains all the characteristic fluid properties. In this case the Hedstrm number is used to calculate the ratio of the inserted material properties and the material properties of the layer. The dimensionless number mentioned determine the ratio between width, length and height of the clean silt layer and the missing dimensions were based on geometrical scaling. This resulted in a test stand which was built in the Laboratory of Soil Movement of the Delft University of Technology.

THE TEST STAND.

The test stand consists of two main parts, a glass aquarium and an insertion pump. The aquarium is filled with a layer of hair gel with a height 0f 0.3 m, simulating the clean silt layer. Figure 4 gives an impression of the aquarium used.

The insertion pipe is mounted on a supporting frame on top of the aquarium.

The second main part of the test stand is the insertion pump (see figure 5).

This pump, developed for this research is based on the principle of a piston pump. According to this principle it is possible to insert material with a constant and well defined volume rate. The insertion pump is driven by an electronically speed controlled handdrill. The maximum volume which can be inserted in one stroke is 5.4 liters.

During the preparations for each experiment it came out to be very difficult to keep the insertion system free of enclosed air bubbles.

The insertion pump is connected with the insertion pipe by a plastic hose. During the experiments 3 differently shaped insertion pipes, with respect to the insertion mouths as shown in figure 6.

In the production process of the hair gel, air bubbles are caught in the substance because of the yield stress. The presence of air bubbles obstruct the transparency of the material because the light which shines from the back through the aquarium diffuses strongly and thus the penetration of the light in the transparent hair gel is not sufficient. The appearance of those air bubbles has an effect similar to fog. Placing the gel in a high vacuum of 5kPa absolute pressure for some time, will cause the gel to boil. This way it is possible to remove the air bubbles.

To distinguish the inserted material from the transparent material, the inserted material is dyed. Addition of dye appeared to be of no influence on the physical properties of the substance. Behind the aquarium a light source is mounted. This light penetrates the transparent gel, but does not penetrates the dyed inserted gel. This makes it possible to see the boundary surface and thus the shape of the dyed inserted gel body.

The gel which is used for the experiments is on waterbasis. The desired gel strength can be obtained by diluting gel with water.

The physical properties of the different dilutions of gel with water can be measured by testing the substance in a roto-viscometer. In this rotoviscometer a sample is subjected to a range of well defined velocity gradients. The shear stress occurring at the different velocity gradients are measured. As a result of these measurements it is possible to make a graph depicting the shear stress as a function of the velocity gradient. An example of such a graph and the graph of the derived dynamical viscosity as function of the velocity gradient, both for undiluted hair gel are depicted in figure 7.

It appears that shear stress as a function of the velocity gradient mathematically can be described as a Casson fluid model. In case of the undiluted gel as depicted in figure 7 the substance can be described as:

(10)

TESTING PROGRAMME.

As stated in the paragraph concerning dimensionless numbers, the strength ratio between the layer material and the inserted material may have different values. As a consequence of this, in the model tests different values of this ratio should also be investigated. Because the injected material can be stronger or weaker than the layer material, for the layer material the gel is diluted in the ratio of 60% gel and 40% water. The advantage of this dilution is that tests can be carried out by inserting material of a stronger as well as a weaker dilution.

This way it was possible to do tests by inserting material with a shear strength varying between 0.25 and 2.5 times the strength of the layer material.

Other parameters which have been varied are, the insertion velocity and the type and shape of the insertion mouths. The tests and the parameters used are given in table 1.

The first 9 tests were carried out with respect to the feasibility study.

The first 9 tests could also be used with respect to the sensitivity study but were supplemented with 5 additional tests.

The first 9 tests included tests with 3 different insertion flows and 3 differently shaped insertion mouths. The dilution of the inserted material was fixed at 30% with a yield stress of 8 Pa.

For the sensitivity study tests were carried out with 3 different dilution percentages. So a total of 4 different ratios of material properties of the inserted gel and transparent gel.

Nr. Test	dilution inserted gel [%]	yield stress [Pa]	type of insertion mouth	insertion flow [cm^3/s]
1	30			83
2	30	8	3-D disc	167
3	30			375
4	30			83
5	30	8	horizontal	167
6	30		one direction	375
7	30			83
8	30	8	vertical	167
9	30			375
10	15	4	3-D disc	83
11	15		vertical	83
12	100	40		83
13	60	16	3-D disc	83
14	60		vertical	83
Table 1: Test programme.

RESULTS.

From the tests carried out, a number of phenomena can be distinguished. The first nine tests were carried out with a dilution of 30% gel and 70% water as inserted material. All these tests resulted in a more or less globular shaped body of the inserted material. In the case of the 3-dimensional disc mouth the body develops as a torus. A number of consequent stages of the development of the inserted dyed gel body are as illustrated in figure 8.

The three different velocities used did not have much influence on the shape of the formed body. The globular body only stretched a bit in the direction of the velocity. The globular shape was not much affected by large air bubbles which were escaping from the insertion mouth through the body to the layer surface. This indicates a stable situation.

The vertical insertion mouth gave the best approach to a perfect global shape of the body.

It appeared that the insertion process could be reversed in such a way that almost no dyed gel remained in the layer of clean gel after the suction process was completed. This is a strong indication that no shear stresses are mobilized on the radial and tangential surfaces inside the inserted silt body.

The tests 10 and 11 gave a totally different result from that of the preceding tests. In these tests a fairly weak dilution of 15% gel with 85% water was inserted. This resulted in inserted material breaking out to the surface of the layer after the insertion started. This phenomena was initiated by a large air bubble enclosed in the inserted material and escaping to the surface. As the insertion process proceeds, the inserted material flows directly to the surface through the channel generated by the air bubble. A scheme of this mechanism is illustrated in figure 9.

When the above described process occurs, it is not possible to suck up all the inserted material. This means the process is not reversible.

In test 12 a 100% concentration of gel was inserted in the transparent gel of 60% concentration. So, the inserted material is stronger than the layer material. The result was a very stable globular body. This shape is not disturbed at all by air bubbles which escape to the surface. In the test it was also possible to suck up the inserted material.

In the experiments 13 and 14, the both gels were of the same concentration. In these cases also a stable globular body developed. The process could also be reversed with these tests.

CONCLUSIONS.

The conclusion which can be drawn from the feasibility study is that it should be possible to form stable torus shaped or globular shaped bodies when inserting a Casson fluid in another Casson fluid with possibly different reological properties.

From the sensitivity study the conclusion can be drawn that two mechanisms may occur. If the material inserted has the same strength as the receiving material a globular body is formed. The stronger the inserted material is, the more stable this mechanism occurs. In principle the developed globular shaped body can be sucked up from the layer meaning the inserted material can be removed again. The type of insertion mouth is of minor influence for the insertion.

If the inserted material is weaker, still a globular body is formed. This globular shape is getting less stable as the material gets weaker. As soon as the strength of the inserted material exceeds a certain boundary value this material appears to break out to the free surface. When this mechanism occurs, it is not possible to suck up all the inserted material. This means the process is not reversible.

The research carried out is promising with respect to the storage of contaminated silt in a layer of clean silt.

Although the results are promising, additional research is necessary to determine the circumstances under which silt bodies are formed in practice within the context of the problem described in this paper.

In prototype a solitary globular shaped body is restricted to a maximum volume.

For practical applications, the contaminated silt could possibly be stored in a layer of clean silt by creating multiple globular bodies. A possible method of application of this discrete injection method is illustrated in figure 10.

Another injection model could be starting the injection with one globular shaped body and continuing the injection while moving the vessel at a predetermined velocity, creating a rounded cylinder shaped body (see figure 11).

Hopefully this research may provide a contribution to the solution of the environmental problems of contaminated silt.

LIST OF SYMBOLS USED.

Kd	distribution coefficient solid-liquid	ml/g
Ca	concentration contaminants adsorped in dry soil	g/g
Co	concentration contaminants dissolved in water	g/ml
dv/dy	velocity gradient	l/s
t	shear stress	Pa
t0	yield stress	Pa
hi	dynamical initial viscosity	Pas
hB	dynamical Bingham viscosity	Pas
ha	dynamic apparent viscosity	Pas
hN	dynamic Newtonian viscosity	Pas
U	internal energy	J
l	efficiency coefficient	-
Qw	heath	J
W	energy	J
Ekin	kinetic energy	J
Epot	potential energy	J
m	mass	kg
cp	specific heath capacity	J/kgK
T	temperature	K
V	volume	m^3
p	pressure	Pa
v	velocity	m/s
r	density	kg/m^3
Re	Reynolds number	-
He	Hedstrm number	-

BIBLIOGRAPHY.

[ 1] Brauer Heinz, Grundlagen der Einphasen- und Mehrphasen stromungen.

Sauerlnder A.G., Arau (Schweiz) 1971.

[ 2] Davids S.W., Berging verontreinigd slib. T.U. Delft 1991.

[ 3] Davids S.W., Een modelonderzoek naar het inbrengen van een Binghamse

vloeistof in een tweede Binghamse vloeistof met andere reologische eigenschappen. T.U. Delft 1991.

[ 4] Exlog staff, Theory and applications of drilling fluid hydraulics. D. Reidel Publishing Company 1984.

[ 5] Harris, J., Rheology and non-Newtonian flow. Longman Group Limited, London 1977. ISBN 0-582 463319.

[ 6] Langhaar, Henry L., Dimensional Analysis and theory of models. Robert E. Krieger Publishing Company.

[ 7] Makkink, A.D.J., Het opslaan van verontreinigde baggerspecie in de IJmeer zandwinput en het ontwerp van de stortinrichting. T.U. Delft 1989.

[ 8] Verhoeven et al, F.A., Boskalis b.v. The relevant soil properties for dredging silt. Proc. Wodcon XI, Brighton U.K. 1986.

[ 9] Koning, J. de & Miedema, S.A., Disposal of contaminated silt in the IJmeer. Ass. Of student Makkink, Delft University of Technology, 1989.

[10] Koning, J. de & Miedema, S.A., Feasibility study of a silt injection method. Ass. of student Davids, Delft University of Technology, 1991.

[11] Miedema, S.A., PLOSIM V4.00. Data processing and presentation software. Delft, Holland, 1991.