UFL/COEL2000/011
PERFORMANCE REPORT ON THE SIMULATOR OF EROSION
RATE FUNCTION
by
Neil K. Ganju
Kevin M. Barry
and
Ashish J. Mehta
September, 2000
Submitted to:
U.S. Army Engineering Research and Development Center
Waterways Experiment Station
3909 Halls Ferry Road
Vicksburg, MS 39180
REPORT DOCUMENTATION PAGE Form Approved
OMB No. 07040188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathenng
and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of
information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite
1204, Arlington, VA 222024302, and to the Office of Management and Budget, Paperwork Reduction Project (07040188), Washington, DC 20503.
1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
September, 2000 Final Report 2/12/99 5/15/00
4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
PERFORMANCE REPORT ON THE SIMULATOR OF EROSION
RATE FUNCTION DACW3999P0238
RATE FUNCTION
6.AUTHOR(S) Neil K. Ganju UF# 451136312
Kevin M. Barry
Ashish J. Mehta
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESSES) 8. PERFORMING ORGANIZATION
Coastal and Oceanographic Engineering Program REPORT NUMBER
Civil and Coastal Engineering Program UFL/COEL2000/011
University of Florida
Gainesville, FL 32611
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESSES) 10. SPONSORING/MONITORING
AGENCY REPORT NUMBERS
U.S. Army Engineering Research and Development Center AGENCYREPORTNUMBERS
Waterways Experiment Station
3909 Halls Ferry Road
Vicksburg, MS 39180
1I. SUPPLEMENTARY NOTES
12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE
13. ABSTRACT (Maximum 200 words) Numerous rotating cylinder devices have been constructed to simulate the erosion characteristics of
cohesive sediments. Typically, these devices consist of a rotating outer cylinder, which is filled with the chosen eroding fluid. A remolded or
intact sample in the shape of a smaller cylinder is placed into the outer cylinder. While rotating, the outer cylinder imparts rotation to the fluid,
which translates into a shear stress being uniformly imparted to the sample. The torque applied to the sample is measured and converted to the
corresponding shear stress. Erosion of the sample follows, and a relation is obtained between shear stress and erosion rate. A rotating cylinder
device the Simulator of Erosion Rate Function (SERF) was constructed and it's performance tested. SERF consists of an rotating acrylic
outer cylinder filled with eroding fluid, and an inner stationary cylinder which holds the sample between brackets. A torque cell is connected
to the stationary cylinder via a shaft, and measures the applied torque. A load cell connected in turn to the torque cell measures the weight of
the sample. SERF was initially tested with aluminum cylinders representing dummy samples to calibrate for shear stress, following which three
commercial types of ceramic clay samples were tested. Each sample was run by stepincreasing the rpm and the loss in sample mass was
recorded over each timestep. A relation (erosion rate function) between erosion rate and shear stress was obtained for the sample. The erosion
rate function was found to be linear or piecewise linear, depending on the degree of uniformity of the sample. Furthermore, increasing shear
strength correlated with increasing density for all three clay types.
14. SUBJECT TERMS 15. NUMBER OF PAGES
31
16. PRICE CODE
17. SECURITY 18. SECURITY 19. SECURITY 20. LIMITATION OF ABSTRACT
CLASSIFICATION OF REPORT CLASSIFICATION OF THIS CLASSIFICATION OF
PAGE ABSTRACT
Standard Form 298 (Rev. 289)
Prescribed by ANSI St. Z3918
298102
NSN 7540012805500
UFL/COEL2000/011
PERFORMANCE REPORT ON THE
SIMULATOR OF EROSION RATE FUNCTION
By
Neil K. Ganju
Kevin M. Barry
and
Ashish J. Mehta
Submitted to:
U.S. Army Engineering Research and Development Center
Waterways Experiment Station
3909 Halls Ferry Road, Vicksburg, MS 39180
Coastal and Oceanographic Engineering Program
Civil and Coastal Engineering Department
University of Florida
Gainesville, FL 32611
September, 2000
SUMMARY
Numerous rotating cylinder devices have been constructed to simulate the erosion
characteristics of cohesive sediments. Typically, these devices consist of a rotating outer
cylinder, which is filled with the chosen eroding fluid. A remolded or intact sample in the shape
of a smaller cylinder is placed into the outer cylinder. While rotating, the outer cylinder imparts
rotation to the fluid, which translates into a shear stress being uniformly imparted to the sample.
The torque applied to the sample is measured and converted to the corresponding shear stress.
Erosion of the sample follows, and a relation is obtained between shear stress and erosion rate. A
rotating cylinder device the Simulator of Erosion Rate Function (SERF) was constructed and
it's performance tested. SERF consists of an rotating acrylic outer cylinder filled with eroding
fluid, and an inner stationary cylinder which holds the sample between brackets. A torque cell is
connected to the stationary cylinder via a shaft, and measures the applied torque. A load cell
connected in turn to the torque cell measures the weight of the sample.
SERF was initially tested with aluminum cylinders representing dummy samples to
calibrate for shear stress, following which three commercial types of ceramic clay samples were
tested. Each sample was run by stepincreasing the rpm and the loss in sample mass was
recorded over each timestep. A relation (erosion rate function) between erosion rate and shear
stress was obtained for the sample. The erosion rate function was found to be linear or piece
wise linear, depending on the degree of uniformity of the sample. Furthermore, increasing shear
strength correlated with increasing density for all three clay types.
This work was supported by funds from The U.S. Army Engineering Research and
Development Center, Waterways Experiment Station, Vicksburg, MS, through contract
DACW3999P0238.
TABLE OF CONTENTS
SU M M A R Y ............................................................. ..... ....................... ......... 2
LIST OF FIGURES.................................................................... ................. 4
LIST OF TABLES................. .... ............ .. .... .... ........................................
LIST OF SYMBOLS.......................................................... ..... ... ..........6
SECTIONS
1. Introduction ...................................................... .................. ............. .. 8
2. SE R F C haracteristics........................................................ .............. ..... ........... 8
3. SERF Calibration ................................................................................ ..........
4. SERF Testing..................................................................... .........11
5. SERF Results ................................................... ........................ ....... 13
6. Concluding Com m ents.................................................................. ........... 15
7. References.................................................................... ...... ............. 17
APPENDIX USER'S MANUAL FOR SERF.........................................................19
A .1. SERF Calibration ....... ....... ............................................................... ..... .....19
A 1.1 Torque Cell Calibration..................................................................19
A. 1.2 Load Cell Calibration............... ...... ................................... ...19
A.1.3 Shear Stress Calibration ....... ... .............................................20
A.2. Sample Preparation.............. ...............................................................22
A .3. Sam ple Testing..................... ...... ........... .. ................. ................... .............. 23
LIST OF FIGURES
2.1 Schematic of SERF assembly..................................................................24
2.2a Frontview of SERF...................................................... ....... ......... .....24
2.2b Photograph of acrylic outer cylinder and measurement cells...... .............................25
3.1 Calibration curve for large aluminum cylinder (dia. = 7.6 cm).............................25
3.2 Calibration curve for medium aluminum cylinder (dia. = 7.2 cm).............................26
3.3 Calibration curve for small aluminum cylinder (dia. = 6.8 cm)...............................26
3.4 Calibration curve for bottom disk................... .... .. .. ............. ........27
5.1 Erosion rate as a function of shear stress for clay type #10.................... ......... 27
5.2 Erosion rate as a function of shear stress for clay type #20 ..................................28
5.3 Erosion rate as a function of shear stress for clay type #75................................28
5.4 Erosion rate as a function of shear stress for clay type #10, with differing pore
fluid (tap water vs. saline water)..................................................................29
5.5 Erosion rate as a function of shear stress for clay type #20, with differing pore
fluid (tap water vs. saline water)...................................................... ....... 29
5.6 Erosion rate as a function of shear stress for clay type #75, with differing pore
fluid (tap w ater vs. saline water)............................... .......... ........... ..............30
5.7 Erosion rate constant as a function of shear stress for clay type #75........................30
5.8 Shear strength as a function of dry density for three clays...................................31
LIST OF TABLES
5.1 Density, erosion rate constant, and shear strength values obtained for 15 clay
sam p les .................................... .................................................. ....... 14
LIST OF SYMBOLS
A surface area (m2)
FT applied tangential force (kg)
R inner cylinder radius (m)
Re Reynolds number ()
RPM revolutions per minute
T torque (kgm, gcm)
Ta Taylor number ()
U annular peripheral velocity (m/s)
d annular gap width (m)
g acceleration due to gravity (m/s2)
Ams change in dry sediment mass (kg)
r radius (m)
At time interval (s)
a empirical coefficient, Eq. 5.3 ()
/f empirical coefficient, Eq. 5.3 ()
Xs empirical coefficient, Eq. 5.2 ()
e erosion rate (kg/m2s)
EN erosion rate constant (kg/Ns)
NO maximum erosion rate constant (kg/Ns)
As empirical coefficient, Eq. 5.2 ()
p sediment bulk density (kg/m3)
Pd dry density (kg/m3)
Pf pore fluid density (kg/m3)
Pt fluid mud density (kg/m3)
Ps sediment granular density (kg/m3)
r shear stress (Pa)
Ib sample shear stress (Pa)
Ts shear strength (Pa)
v kinematic viscosity (m2/s)
PERFORMANCE REPORT ON THE
SIMULATOR OF EROSION RATE FUNCTION
1. Introduction
Numerous rotating cylinder devices have been constructed to simulate the erosion
characteristics of cohesive sediment (Moore and Masch, 1962; Arulanandan et al., 1975; Croad,
1981; Chapuis and Gatien, 1986). Lee and Mehta (1994) have provided a brief summary of the
underlying designandoperation concept. Typically these devices consist of a rotating outer
cylinder, which is filled with the chosen eroding fluid A remolded or intact sample in the shape
of a smaller cylinder is placed into the outer cylinder. While rotating, the outer cylinder imparts
rotation to the fluid, which translates into a shear stress being uniformly imparted to the sample.
The torque applied to the stationary sample is measured and converted to the corresponding
shear stress. Erosion of the sample follows, and a relation (erosion rate function) is obtained
between erosion rate and shear stress. The device requires calibration to convert the torque into
an applied shear stress, as well as some method to measure the eroded mass. For the unit
developed and tested, i.e., the Simulator of Erosion Rate Function serfF), dummy aluminum
cylinders were used to calibrate for the shear stress via the attached torque cell. A load cell
allowed the eroded mass to be measured without removing the sample assembly from the outer
cylinder. The inclusion of the load cell is an improvement over previous devices, which required
removing either the sample or fluid to measure eroded mass. Three types of commercial clays
were tested to gauge the performance of the SERF.
2. SERF Characteristics
Figure 2.1 shows the schematic of the SERF, and Figs. 2.2a,b show photographic views.
The clay sample is molded into a cylinder with a diameter of 7.6 cm and a height of 9.6 cm. The
cylinder is impaled on a shaft which secures an upper and a lower disk of equal diameters to the
two ends of the sample. The upper end of the shaft affixes to an Omega (Stamford, CT) model
TQ201100Z torque cell, which is in turn connected to an Omega model LC6015 load cell.
When the acrylic outer cylinder is rotated, the chosen eroding fluid imparts a torque to the
sample, and in turn the shaft and torque cell. The cell is connected to an Omega model DP25S
LED (lightemitting diode) display, and shows the applied torque in gcm. The load cell is
attached to the entire shaft/torque cell/sample assembly, and can be used to measure mass eroded
after a given run. This unit can move vertically out of the outer cylinder for sample and water
replacements.
The outer cylinder is capable of rotating at a maximum of 2,350 rpm, and a minimum of
150 rpm. The torque cell has a 1,800 gcm capability, and the load cell can accommodate a
maximum of 2.3 kg.
3. SERF Calibration
The torque cell is calibrated using an aluminum ring which affixes to the shaft of the cell.
A twopoint calibration is done using no load, and then with a 1,000 g weight hanging from the
ring, which has a radius of 1 cm. This results in a calibration between 0 and 1,000 gcm. The
load cell is calibrated in the same twopoint fashion, using a 1,000 g weight hung from the cell.
The calibration to convert torque into shear stress requires running the device with
dummy samples. The SERF was run with three aluminum cylinders of varying diameter and
identical height. The varying of diameters provides a relationship between the applied torque and
the sample radius, considered to be necessary in instances when the radius of a sample decreases
measurably during testing. For a given cylinder, the SERF was rotated from a standstill to a
"reasonable" speed (up to 1,600 rpm) in timestep increments, and torque readings were taken at
each rpm value, after steadystate was reached (typically within 35 seconds). Once at peak rpm,
the device was slowed in timestep increments and torque readings were taken again. This
procedure was done three times for each cylinder, the shaft and cylinder being removed and re
installed each time in order to gauge the precision of the device after reinstallment of samples.
The bottom disk of the sample holder was also tested in this manner. By subtracting the bottom
disk torque from the total torque, the torque applied to the sample area only can be obtained. The
meniscus must be located just below the upper disk (Fig. 2.1), so the fluid is acting on the
cylinder and the bottom disk. It is sometimes necessary to add or remove water to obtain an
appropriate meniscus.
A trendline was plotted for the variation of torque with rpm for the disk, and the
corresponding relation (Eq. 3.1) was used to determine the torque provided by only the sample
area of the cylinder, by subtracting the bottom disk torque (Fig. 3.4) from the overall torque. This
adjusted torque value was converted into shear stress using Eqs. 3.2 and 3.3:
T= 0.0086RPM 1 (3.1)
Fr= T (3.2)
r
=gFr (3.3)
A
where T is the torque reading in kgm (gcm for Eq. 3.1), r is the cylinder radius in m, FT is the
applied tangential force in kg, r is the shear stress in Pa, g is the acceleration due to gravity in
m/s2, and A is the cylinder surface area in m2. The shear stress values were then plotted versus
rpm to yield calibration curves for three cylinders (Figs. 3.1, 3.2, 3.3). The spread in data is
attributed to slight, inevitable changes in the alignment of the cylinders with respect to the
central axis, due to the removal and reinstallment of the shaft/cylinder assembly between runs.
This procedure of removal/reinstallment was carried out in order to identify the degree of
precision of the device. Overall, the small and large cylinders followed a power fit, and the
medium cylinder followed a linear trend (Figs. 3.1, 3.2, 3.3). Fitting a power function to the
medium cylinder resulted in overestimation of shear stress at high rpm (>1,000).
For the large, medium, and small cylinder, respectively, the following relations were
obtained:
t = 4.9 x 106 RPM96 (3.4)
z = 4.7 x 103RPM 1.32 (3.5)
t = 8.0 x 108 RPM2.63 (3.6)
As an illustration of the differences among the above relations, note that a 10% decrease in
surface area from the small cylinder to the large cylinder leads to a 67% increase in shear stress
at an rpm of 1,000 for the small cylinder. This results from the decreased moment arm of the
small cylinder leading to an increased tangential force and therefore increased shear stress.
4. SERF Testing
Once the relation of shear stress to rpm was determined, clay samples could be tested for
their erosion rate as a function of applied shear stress. Three ceramic clays were obtained from
Bennett Pottery Supply (Orlando, FL), #10, #20, and #75. The company did not provide any
information concerning the compositions of these clays. In any event, each clay was molded to
two different bulk densities (both lower than initial) by adding tap water and working by hand.
This procedure was repeated twice, resulting in four densities tested for each clay. A third set
was prepared by molding two equal amounts of clay with water, one with tap water and another
with 35 ppt salt (sodium chloride) water. Water was added in equal amounts to yield a similar
density. Since the clays arrived with existing pore water, the actual salinity of the pore water
was less than 35 ppt.
The first set of clays was shaped into cylinders of the same diameter as the medium
cylinder, hence the medium cylinder calibration curve was used. The second set of samples was
shaped into the diameter of the large cylinder, as was the third set. The volume and surface area
of each sample was known, and the mass of the sample was determined by weighing the sample
with the bracket and lid (Figs. 2.1, 2.2b), and weighing only the bracket and lid. This provided
the density of each respective sample.
Once a sample was prepared it was lowered into the outer cylinder, and the load cell was
zeroed. After 5 to 15 minutes of rotation at a given rpm, the device was stopped and allowed to
rest for approximately one minute. The timestep was chosen based on observation of the
erodibility of a given sample during the previous timestep. In most cases the maximum speed
was 1,500 rpm. Above this speed, the possibility of mass erosion (i.e., large clasts being eroded)
was present, which could have resulted in an imbalance and excessive torque, thereby damaging
the torque cell. After reaching steadystate, the load cell displayed the loss in material in grams.
The load cell displayed the loss of sediment as well as the internal pore fluid which was
also released. Since the load cell is measured buoyant mass, only the mass loss of solids was
being displayed. If the pore fluid were assumed to be of similar density as the eroding fluid, then
the load cell displayed the loss of dry sediment mass. This value, along with the timestep and
the surface area resulted in the erosion rate:
Ams
E= S (4.2)
AAt
where Ams is the dry sediment mass loss, A is the sample area in m2, and At is the time interval in
seconds.
5. SERF Results
Eighteen clay samples were tested in total, and erosion rate as a function of shear stress
was obtained (Figs. 5.1 through 5.6) for all but three samples (#2, 9, and 15) whose erosion was
undetected by the load cell. Linear (or piecewise linear) trendlines were fitted to each set of
results with the following form:
E= EN (rb s) (5.1)
where is the erosion rate in kg/m2s, r, is the sample shear strength in Pa, EN is the erosion rate
constant in kg/Ns (=e when rb = 2,r), and Tb is the shear stress, also in Pa Table 5.1).
The composite plots (Figs. 5.1 through 5.6) for each clay show a general trend towards
increasing shear strength with increasing density for all three clays. For the case of saline pore
fluid versus fresh pore fluid, the saline samples tended to have greater erosion rates at the same
shear stress than the fresh samples, thus implying a loss of soil strength due to the addition of
sodium chloride. Unfortunately, because clay compositions were not available, an explanation
for this effect of sodium chloride cannot be deduced. Some of the results show a twoline trend
as opposed to a singleline trend. This Type 1/Type 2 erosion phenomenon (Mehta, 1981)
possibly results from the nonuniformity of the sample, whereby the erosion rates are not
constant between different density zones of the sample. The handmolding of the clay samples
likely left a surface layer of nonuniform and possibly lower density than the interior of the
sample in these cases.
Table 5.1. Density, erosion rate constant, and shear strength values obtained for 15 clay samples
Sample Density, p EN (kg/m2s) Ts (Pa)
#, type (kg/m3)
1, 10f 1665 5.00 x 105 3.14
3, 10f 1710 1.91 x 105 3.38
4, 10f 1859 1.23 x 105 4.49
5, 10s 1928 1.73 x 105 0.60
6, 10f 1928 1.20 x 105 1.67
7, 20f 1435 3.93 x 105 1.93
8, 20f 1537 1.19 x 105 2.81
10, 20f 1721 7.72 x 10'5 3.02
1.73 x 105,
11, 20s 1894 x 0 1.58
7.00 x 10'
12, 20f 1905 7.10 x 106 2.00
13, 75f 1675 4.80 x 105 1.39
14, 75f 1698 3.217 x 105 1.73
16, 75f 1806 1.14 x 105 1.76
3.20 x 10',
17, 75s 1940 6.89 x 10 3.27
18, 75f 1963 9.85 x 106 1.88
Note: s indicates saline pore fluid, f indicates tap water pore fluid
The erosion rate constant is plotted versus shear strength (Fig. 5.7) for one set of tests
with clay #75, and the trend in this case follows the fit provided by Mehta and Parchure (2000).
The equation contains two empirical coefficients and is represented as
SN = sNO exp (Xszr) (5.2)
where ENO is selected to be 0.2 kg/Ns based on the analysis of Mehta and Parchure, and Zs = 7.45
and As = 0.5 are obtained as bestfit values.
The shear strength of each sample was plotted versus the dry density, and a fit of the
following form (Mehta and Parchure, 2000) was applied to each clay type (#10, #20, #75):
rs = a(pdplf (5.3)
where a and p are empirical coefficients, pd is the dry density in kg/m3, and pi is 64 kg/m3
(selected to represent the dry density of fluid mud with no shear strength). The dry density was
calculated from the mass balance
p= P (5.4)
For the #10 and #20 clays, a= 3.20 x 103, and l=1, and for #75, a= 1.30 x 103, and P=1 were
obtained (Fig. 5.8).
It should be pointed out that the selected group of samples did not erode enough to
significantly decrease the sample radius, and only one cylinder calibration curve was used for
each sample. For example, sample #6, with a bulk density (wet) of 1928 kg/m3, lost a total of 3 g
during the entire test. This corresponds to a reduction in thickness of 0.007 cm, or 0.18%
thickness lost (based on large cylinder volume).
6. Concluding Comments
A comment is in order concerning the operating speed range of SERF, the generation of
Taylor vortices in the annular gap between the outer cylinder and the sample, and the
development of turbulence in this gap. When a certain Reynolds number has been exceeded in
the gap between two rotating concentric cylinders, vortices appear in the flow. The axes of these
vortices are along the circumferences of the cylinders, and they rotate in alternately opposite
directions (Schlichting, 1968). The Taylor number is used to predict the formation of these
irregularities, and is defined as
= dd (6.1)
v R
where U is the peripheral velocity of the cylinder, d is the gap width, R is the inner cylinder's
radius, and v is the kinematic viscosity of the fluid. Taylor vortices are formed in the range
41.3
speed of SERF, which corresponds to a peripheral velocity of 0.8 m/s (at 150 rpm), using a gap
width of 0.024 m, an inner cylinder radius of 0.038 m, and water as the fluid (V106 m2/s), the
Taylor number is 15,000, well over the range of Taylor vortex formation.
With the characteristic Reynolds number defined as
Re =Ud (6.2)
v
using a peripheral velocity of 0.8 m/s and the same fluid and gap width, Re=18,880 is obtained.
The SERF is typically operated above this speed, and in these tests the lowest Re attained was
100,000 at a speed of 4.3 m/s (at 800 rpm). Operating above the Taylor vortex regime and above
the laminar flow regime, one can expect a comparatively uniform distribution of shear stress
over the sample surface area.
The SERF is capable of producing coherent results on the rate of sample erosion as a
function of applied shear stress, even with small amounts of erosion (tenths of a gram). It should
be feasible to use the apparatus to test a variety of cohesive samples, both intact from the field
and remolded. The inclusion of a load cell directly attached to the apparatus considerably
reduces the time necessary to obtain results, and ensures accuracy by not requiring periodic
removal and weighing of the sample assembly. The device is operated at sufficiently high
Reynolds numbers (>100,000 at 800 rpm) to ensure a comparatively uniform shear stress
distribution.
With regard to the clay samples tested the following conclusions can be drawn:
1. Linear and piecewise linear erosion rate functions were observed for three types of
commercial clays,
2. The clays displayed a general trend toward increasing shear strength with density,
3. The use of saline pore fluid resulted in a greater erosion rate constant than samples
prepared with tap water.
A user's manual for SERF is included as Appendix.
7. References
Arulanandan, K., Loganathan, P., and Krone, R.B., 1975. Pore and eroding fluid influences on
surface erosion of soil. Journal of Geotechnical Engineering Division, ASCE 101(1), 51
66.
Chapuis, R.P., and Gatien, T., 1986. An improved rotating cylinder technique for quantitative
measurement of the scour resistance of clays. Canadian Geotechnical Journal, 23(1), 83
87.
Croad, R.N., 1981. Physics of erosion of cohesive soils. Ph.D. Thesis, University of Auckland,
New Zealand.
Lee, S.C. and Mehta, A.J., 1994. Cohesive sediment erosion. Report DRP9406, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS, 1215.
Mehta, A.J., 1981. A review of erosion functions for cohesive sediment beds. Proceedings of the
First Indian Conference on Ocean Engineering, Vol. 1, Madras, 122130.
Mehta, A.J. and Parchure, T.M., 2000. Surface erosion of finegrained sediment revisited. Muddy
Coast Dynamics and Resource Management, B. W. Flemming, M.T. Delafontaine, and
G. Liebezeit, eds., Elsevier, Oxford, UK (in press).
Moore, W.L., and Masch, F.D., Jr., 1962. Experiments on the scour resistance of cohesive
sediments. Journal of Geophysical Research, 67(4), 14371449.
Schlichting, H., 1968. BoundaryLayer Theory, 6th edition, McGrawHill, New York, 500503.
APPENDIXUSER'S MANUAL FOR SERF
A. 1 SERF Calibration
The SERF requires three calibrations: torque cell calibration, load cell calibration
and shear stress calibration.
A. 1.1 Torque Cell Calibration
Remove the torque cell by unscrewing the single hex screw which connects the
torque cell to the load cell (Fig. 2.1). Use caution in handling the torque cell, which can
be damaged if overtorqued. Secure the torque cell base to a horizontal surface with a
vise, keeping the shaft of the torque cell aligned in the horizontal direction. Place the 1
cm radius ridged collar (provided with the apparatus) around the shaft, and tighten the
collar hex nut so it secures to the shaft without rotating. Secure a loop of string to the
collar; tighten it to prevent rotation. Follow the Omegaprovided calibration instructions
for the lightemitting diode (LED) display (Omega DP25S), applying zero load for the
0% point, and a 1 kg mass for the 100% load. The readout will now display torque in g
cm in the range of 01000 gcm. To ensure proper calibration, secure an intermediate
mass to the string, and the display should show the corresponding torque (500 g load
should result in a 500 gcm torque). Remove the load, string and collar, and set the torque
cell aside.
A.1.2 Load Cell Calibration
Remove the load cell by unscrewing the two vertical hex screws at the rear of the
cell. Secure the load cell horizontally with a vise. Loop a string through the hole at the
end of the cell. Securing the load to the end of the string, follow the provided calibration
instructions for the LED display, applying zero load for the 0% point, and 1 kg for the
100% point. Since SERF is only concerned with small mass changes, the load cell does
not have to be calibrated above 1 kg. If smaller changes in mass are anticipated, the cell
can be calibrated to 500 g or less, and the decimal point can be moved for higher
resolution ( 0.1 gram finest resolution). Decimal point movement is explained in the LED
display manual. Replace the load cell with the two hex screws, and replace the torque
cell, tightening slowly.
A.1.3 Shear Stress Calibration
Attach the chosen aluminum cylinder to the shaft, securing it between the upper
and lower disks/brackets (Fig. 2.1). The lower bracket should be flush with the shaft, and
the upper bracket should tighten onto the cylinder. Screw one of the two hex nuts onto
the upper portion of the shaft, so it tightens against the top of the upper bracket. Tighten
with an adjustable monkey wrench. Screw the other nut onto the shaft, approximately
two cm above the lower nut. Place the plastic lid (Fig. 2.2b) onto the assembly, and screw
the shaft into the torque cell, twisting counterclockwise slowly, until the shaft rests at the
internal roof of the torque cell (The shaft will halt at the roof). Screw the upper nut
counterclockwise so it locks against the torque cell. DO NOT OVERTIGHTEN! The nut
should be tightened until it just stops turning, no further. For maximum care, twist the
torque cell by hand in the opposite (clockwise) direction to minimize the torque being
applied to the cell. By keeping the LED display on, one can observe how close the cell is
to being overtorqued.
Fill the outer cylinder to approximately 1/3 of it's volume with the chosen eroding
fluid (e.g., fresh water). Lower the cylinder/bracket assembly into the outer cylinder by
loosening the four rearfacing adjustment arms. The assembly should be brought to rest
on the square collar on the back of the post. Tighten the adjustment arms. Fill with more
water until the upper bracket is just submerged.
Slowly spin the outer cylinder counterclockwise (forward setting) at less than
400 rpm and observe the meniscus. A level meniscus is desired. Four horizontal screws
(Fig. 2.2b) can be loosened to move the assembly left and right, and four vertical screws
(Fig. 2.2b) can be loosened to move the assembly forward and back. Small alignment
changes tend to affect the meniscus considerably. After the assembly is centered, the
height of the meniscus should be adjusted so at a given rpm the meniscus is just below
the upper bracket while spinning. It is ideal to start with a lower level at low rpm, and to
add water as the rpm increases and the meniscus lowers.
Begin taking readings after setting the torque cell at zero rpm. Increase the speed
in timestep increments, allowing the torque cell to reach steadystate before recording
the torque value. Once at desired peak rpm (2,350 rpm is attainable, below 1,600 rpm is
prudent), reduce the speed in increments and record torque values. Once this procedure is
completed raise the assembly, loosen the shaft upper nut (clockwise), and unscrew the
shaft from the torque cell (clockwise). Reattach the shaft and repeat the above procedure.
Three runs in this manner are sufficient to obtain the mean trend.
To obtain the calibration for the bottom disk, remove all of the water, and repeat
the above procedure, keeping only the bottom disk submerged at every rpm step. After
obtaining a torque versus rpm trendline, at each selected rpm subtract the bottom disk
torque from the cylinder torque. The residual torque is thus adjusted for the added stress
effect of the bottom disk.
By dividing the adjusted torque readings by the radius of the cylinder, the
corresponding tangential force (in kgmultiply by g for N) is obtained. Dividing this
quantity by the surface area (m2) yields the shear stress in Pa. For each cylinder tested,
plot shear stress versus rpm and obtain the required calibration curves by drawing a mean
trendline through the data points.
A.2. Sample Preparation
Whether using an intact field sample or a remolded sample, the sediment must be
in the shape of one of the three cylinder sizes so the obtained calibration curves can be
applied. In addition, the volume of the sample must be known so it's density can be
measured.
Attach the selected aluminum cylinder to the bracket assembly as described in
Section A.1.3, and then remove the bottom disk and the aluminum cylinder. This is
necessary to place the upper bracket at the correct elevation. Impale the sample on the
shaft. Screw the lower bracket to the end of the shaft, until the shaft is flush with the
bottom of the disk. Plastic sheeting, cut to the height and circumference of the cylinders,
is provided with the apparatus to wrap the sample and ensure the sample has the correct
volume. Trim with a putty knife any extra mass before weighing. Place the plastic lid on
the assembly over the top of the shaft. Weigh the entire assembly. The bracket/lid
combination can be weighed later to provide the weight of only the sample. This weight
divided by the volume gives the (wet) bulk density of the prepared sample.
In the same manner mentioned in Section A. 1.3, secure the assembly to the torque
cell.
A.3. Sample Testing
Fill the outer cylinder about 1/3 with the chosen eroding fluid. Lower the
assembly into the outer cylinder to the collar level. Add water under the lid until the
upper bracket is just flush with the water level. In this case the exact water level is not
very important as long as the sample is fully submerged. The shear stress on the sample
will be independent of the water level, as long as the entire sample is being acted on by
the fluid.
Secure the black torque cell wire to the rear post using a strip of tape, so the load
cell is not affected by it's movement. Zero the load cell. Wait until fluctuations are no
more than 0.1 g before starting rotation.
Rotate the outer cylinder at a given rpm for a given timestep. At the end of the
timestep, stop the device. The load cell usually takes one minute to reach a steadystate,
and then it reads the mass lost during that time interval. Using Eqs. 4.1 and 4.2, the
erosion rate can be determined. The rpm value can be converted to a shear stress via the
calibration curves.
Rotating Acrylic Cylinder
10.16 cm ID
Upper Disk
0.476 cm. S.S. Supporting
Shaft
7.61 cm x 10.16 cm
Cohesive Sediment
Sample
Bottom Disk
Variable speed motor drive
Figure 2.1. Schematic of SERF
Figure 2.2a. Frontview of SERF showing the rotating cylinder assembly, speed control panel
(lower left) and digital readouts (lower right) for the torque cell and the load cell
Figure 2.2b Photograph of acrylic outer cylinder and measurement cells supported by a vertical
member
0 200 400 600 800 1000
RPM
Figure 3.1. Calibration curve for large aluminum cylinder (dia.=7.6 cm)
1200 1400
6
S5
(
40
3 
2
1^ *
1
*
0
0 200 400 600 800 1000
RPM
Figure 3.2. Calibration curve for medium aluminum cylinder (dia.=7.2 cm)
6
5
4
0
0
L3
a '
(I,
 2^
0 200 400 600 800
RPM
Figure 3.3. Calibration curve for small aluminum cylinder (dia.=6.8 cm)
1200 1400
1000
0
0 200 400 600 800 1000 1200 1400
RPM
Figure 3.4. Calibration curve for bottom disk
#1 rho=1665 kg/m^3
#3 rho=1710 kg/m^3
S#4 rho=1859 kg/m^3
Oeo'
1.80E04
1.60E04
1.40E04
1.20E04
1.00E04
8.00E05
6.00E05
4.00E05
2.00E05
0.00E+00
0 1 2 3 4 5
Shear Stress (Pa)
6 7 8 9
Figure 5.1. Erosion rate as a function of shear stress for clay type #10
1.40E04
1.20E04
1.00E04
8.00E05
6.00E05
4.00E05
2.00E05
0.00E+00
0 1 2 3 4 5 E
Shear Stress (Pa)
Figure 5.2. Erosion rate as a function of shear stress for clay type #20
U
1 2 3 4 5
Shear Stress (Pa)
*#13 rho=1675 kg/m^3
#14 rho=1698 kg/m^3
#16 rho=1806 kg/m3
Figure 5.3. Erosion rate as a function of shear stress for clay type #75
*#7 rho=1435 kg/mA3
*#8 rho=1537 kg/mA3
X#10 rho=1721 kg/mA3
1.80E04
1.60E04
1.40E04
1.20E04
1.00E04
8.00E05
6.00E05
4.00E05
2.00E05
0.OOE+00
1.20E04
1.00E04
8.00E05
*6.#10 fresh
6.00E05
M 1#10 saline
C
') 4.00E05
L
2.00E05
0.00E+00
0 1 2 3 4 5 6 7 8
Shear Stress (Pa)
Figure 5.4. Erosion rate as a function of shear stress for clay type #10, with differing pore fluid
(tap water vs. saline)
1.20E04
1.00E04
8.00E05
6.00E05
4.00E05
2.00E05
O.OOE+00
* #20 fresh
* #20 saline
0 1 2 3 4 5 6
Shear Stress (Pa)
Figure 5.5. Erosion rate as a function of shear stress for clay type #20, with differing pore fluid
(tap water vs. saline)
2.00E04
1.80E04
1.60E04
1.40E04
1.20E04
1.00E04
8.00E05
6.00E05
4.00E05
2.00E05
0.00E+00
1 2 3 4 5 6 7
Shear Stress (Pa)
Figure 5.6. Erosion rate as a function of shear stress for clay type #75, with differing pore fluid
(tap water vs. saline)
1.20E04
1.00E04
0 b.UUhUb
0 4.00:E05
0
(U
r
0 4.00E05
w
2.00E05
O.OOE+00
1 1.2 1.4 1.6 1.8 2
Shear Strength (Pa)
Figure 5.7. Erosion rate constant as a function of shear strength for clay type #75
*#75 fresh
S#75 saline
5
4
3
,c
0U
0 "
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Dry Density (kg/m3)
Figure 5.8. Shear strength as a function of dry density for three clays
