SLUDGE HANDLING

CHARACTERISTICS
IN PIPED SYSTEMS

BY J. DAVID BOURKE
 SLUDGE HANDLING
CHARACTERISTICS IN PIPED SYSTEMS
ABSTRACT
Depending on the percentage, size and condition of solids; pipe size; and
many other factors, sludge flow may exhibit near-Newtonian fluid characteris-
tics, non Newtonian homogeneous, or hetero-homogeneous characteristics. A
survey of data for sludge flow characteristics in piping shows the diversity of
results expected from the immense possibility of variables. Hazen-Williams or
the Fanning equations may be used for friction loss calculations in piping,
using factors derived from previously published data, as long as the user recog-
nizes the margin of error possible due to the many variations in sludge content.
Vacuum filter cake can be best described in non-Newtonian terminology as
“False-Body” and a useful tool appears to be the Power Law relationship of;
where:
P is friction pressure drop d = diameter of pipe
L is length of pipe n = power law factor
Q is flow rate

INTRODUCTION quence of the complexity of the dispersed that it moves as a sin-

phenomena rather than errors in gle liquid regardless of the ve-
Although flow behavior of investigation. locity; i.e., a homogenous mix-
most liquids has been mathe- The following is a discussion ture. Although this reference is
matically described and empir- of the various fluid classifica- a liquid-solids mixture, it ob-
ically verified, similar knowledge tions, what complicates the flow viously could also be a liquid-
about solid-liquid mixtures is not calculations, how this applies to gas, or a liquid-liquid mixture.
available. Because of the com- sewage sludge, and what can be A discussion of the types of
plex nature of solid-liquid mix- done to predict its behavior in a single phase fluids and their
tures, postulations about their pipe. behaviors would have to start with
flow dynamics are presently be- First, in the terminology of Newtonian fluids such as water,
yond the reach of the most rheologists, is a description of oil, etc. Here, the rate of shear
sophisticated computer programs. the various types of flow be- of the fluid is directly propor-
It has thus become neces- havior that can take place in a tional to the shearing stress. The
sary to gain this knowledge pipe. proportionality constant which
through empiricisms. Even though The initial classification is into we call viscosity, doesn’t change
a substantial amount of single phase and multi-phase with velocity or shearing rate.
data is beginning to accumulate, fluids. A layman’s definition of Sludge under low concentrations
there exist wide variations a single phase fluid would de- will often approach this condi-
among the results of different scribe it as a single liquid or tion.
investigators — more as a conse- a liquid-solids mixture so finely From there we move into the
more commonly encountered
FIGURE 1 phenomena of non-Newtonian
fluids . . . those which display
pseudoplastic, dilatant, thixo-
tropic, or rheopectic properties.
Pseudoplastic and dilatant fluids
are time independent. Pseudo-
plastics drop in apparent vis-
cosity as shear rate or pipe
velocity increases, while dila-
tants increase in viscosity. Dila-
tant phenomena in slurries to be
handled by pipe lines are very
uncommon. Certain very high
solids contents fluids and filter
cakes of calcium carbonate have
shown dilatant properties; but, a
sewage sludge would seldom, if
ever, display those characteristics.
Rheopectic is the time-depen-
dent equivalent of dilatant. Not
only does the apparent viscosity
increase with an increase in
 FIGURE 2
Page 2

types of thixotropic behavior,

Pryce-Jones labeled this phe-
nomenon “faIse-body” (note
Figure 2.)
Probably one of the most con-
venient formulas for the predic-
tion of non-Newtonian flow is the
power law originally proposed
by Ostwald. Whereas Newton’s
law for true fluids stated that
the shear stress varied directly
as the shear rate, the power law
says that for true fluids, pseudo-
plastic, or dilatant fluids the
shear stress varies as the shear
shear rate or pipe velocity; but, In addition to the above mentioned
rate to the “n”th power. The
apparent viscosity also increases phenomena of apparent viscosity
with time, The faster you shear power factor “n” is equal to 1
varying with shear and time, many
it and the longer you shear it, for Newtonian or true fluids;
fluids exhibit a yield stress below
the most viscous it apparently is greater than one for
which no movement occurs.
becomes, as witnessed by the dilatant fluids; and, is less than
amount of force required. They exhibit rigidity as force
one for pseudoplastic or thixotropic
Were sewage sludge truly is applied up to a certain
fluids.
homogenous, thixotropic be- degree and then break down Newton’s Law of
havior could probably best de- exhibiting usually pseudoplastic Fluid Mechanics
scribe its properties. Thixotropic or thixotropic tendencies. They
fluids become less viscous as the may be called yield pseudoplas-
rate of shear (or agitation) is tics, Bingham plastics, false
increased, and also with the bodies, or any other name an
length of time they are agitated
author may care to coin. Vacuum
or sheared. In those instances
where sludge exhibits these ten- filter cake, centrifuge discharge,
dencies, you would find that and other high concentrations of
even though the velocity in the sludge exhibit this characteristic,
pipe was the same at two differ- Although it continues to drop in
ent points, the viscosity of the apparent viscosity with an in-
sludge at one point would be crease in agitation or shear stress,
higher at that moment than at the when at rest a filter cake still ex-
point a mile further down the hibits a rigid shape that requires
line, because it has been sub-
a certain minimum force (though
jected to the shearing stresses of
flow for a lesser time interval. diminished from the original)
This time-dependency of sludge to make it flow again. In his The fluid analysis techniques
viscosity is obviously trouble- attempt to categorize different at Moyno over the years have
some when it comes to predicting been based primarily on this pre-
pipe line specifications, or cept. (Figure 3) Any function
for that matter, laboratory test- FIGURE 3 raised to a power will produce
ing. It rules out valid data ob-
tained through recirculation of
the sludge. Even laboratory tests
performed with uncirculated sludge
are of questionable validity; for,
if the identical sludge
has been transported it is no
longer in the same conditions of
shear rate and time that it would
be subjected to in the system.
The author of a recent techni-
cal paper on the subject des-
cribed that to the best of his
knowledge he was keeping all
conditions constant and his re-
peatability was ± 40%. Such re-
search is not only discouraging,
but deleterious to the acquisi-
tion of further knowledge.
Page 3
a straight line on a log-log graph. pipe for various values of the (figure 6) Remember, that if the
By plotting viscometer readings factor “n”. (figure 5) For a true shear stress is proportional to the
(either rotational or capillary or fluid where n=1, the flow pat- shear rate to the “n” power when
both) on such a curve, one can tern for streamline flow is para- we plot one against the other
extrapolate to the nominal shear bolic with the maximum velocity on log-log curve, we would have
rate in the pipe and pick off an at the center of the pipe double a straight line. If n = 1 or the
apparent viscosity that is used in that of the average velocity. For shear varies directly as the rate
the Fanning equation to estimate a dilatant fluid (where “n” is of shear, the fluid is New-
pipe friction losses. If sufficient greater than 1) the larger the tonian and would have a slope
material is available for a pump number, the closer the velocity of 1 on the curve or a 45 degree
approaches that of the purely angle. For a perfect pseudoplast,
test, the curve may be verified
theoretical perfect dilatant where (n = 0), the line would be hori-
or altered by adding additional
data points. These points are ga- f
n= . That shape is conical ra- zontal, and for a perfect dilatant
f
(n = ), the line would be verti-
thered by running what is termed ther than parabolic and the maxi-
cal. Behn’s curve shows water
a “cavitation curve”. (Figure 4) mum velocity at the center of the
to have its expected 45 degree
Using Newtonian fluids of vary- pipe is three times the average.
slope; and, sludge of 2.8% solids
ing viscosities such as glucose or As mentioned previously this dil-
is parallel to it, thereby showing
silicone oils, it is simple to de- atant or rheophetic phenomena
Newtonian tendencies or a con-
velop characteristic curves of (“n” greater than 1) does not exist
stant (although higher) viscosity
the point of deviation from the in any known sewage sludge flow over a wide shear rate. On this
straight line capacity vs speed conditions. On the other hand, particular sludge a constant multi-
curve and shape of the curve the perfect pseudoplast, where plier to water curves regardless
for each true viscosity at a given n=0, would flow as a solid of pipe sizes and flow rates, is
N.P.S.H. Obviously the higher the plug, with no difference in ve- valid for streamline flow up to
viscosity the lower the speed locity from the pipe walls to the 3%. However, what happens as
of cavitation. When a non-New- center. Even though a perfect the percentage of solids goes
tonian fluid is run under the pseudoplast could not possibly over 5%. The line is no longer
same conditions, it is usually exist, note that for a pseudo- parallel to water, indicating that
found that it matches none of plastic or thixotropic slurry with the flow is no longer Newtonian
the Newtonian curves, but be- a power factor “n” equal to 1/3, and that a constant multiplier to
haves like one viscosity at speed there is no appreciable difference water curves would no longer
A, another viscosity at speed B, in velocity over the inner third be accurate. In fact, not only
and a third viscosity at speed C. of the pipe, so the center por- is the line not parallel; but, it
Knowing the average shear rate tion flows as a plug. is no longer straight. Calculating
in the pump at those speeds, we Moving back into sewage the “n” factor at various points
then have additional points to af- sludge, how would the various for 10.5% solids shows a range
firm or contradict the slurry’s concentrations fit into any of of almost three to one over the
adherence to the power law; these categories? Vaughn Behn range of velocities or shear rate.
and, additional confidence that published one set of curves ran So, even the convenient power law
the pipe friction calculations will with a capillary viscometer that, calculations do not fit unless
effect a practical degree of ac- although it is obviously going to used over a relatively narrow
curacy. (Note Figure 4) differ with results from other range of pipe velocities.
Worth showing at this time, types of sludges, shows at a This really wouldn’t surprise
is the velocity profile within a glance some discouraging facts. anyone with any knowledge of
slurry flow; because, as previous-
FIGURE 4 ly mentioned, the relationships
really exist for only homogene-
ous or single phase flows, i.e.,
slurries that flow at a continuous
mass velocity without tendencies to
separate or flow at different rates
from that of the carrier fluid. This
usually exists only when the size of
the particles is in the extremely small
low micron range. Obviously, sewage
sludge doesn’t fit this descrip-
tion. Although there is a portion
of the sludge that is fine enough
to flow as a homogeneous slurry,
there are obviously enough large
particles to throw it into the wild,
wild world of two phase or
heterogenous flow. When you
ponder the wide variation in solids
density, shape, and size that would
normally exist in a sewage sludge,
you don’t wonder whether its flow is
following the power law, but whether
 Page 4

you stop to consider that parti-

cles are lagging, or leading the
fluid, migrating into more con-
centrated areas, etc., you realize
that even the percentage of solids
is not constant at every point in
your pipe line. In addition, con-
sider the necessity of calculating
the energy required to move at
certain velocities those particles
not flowing with the liquid, but
tumbling along either a station-
ary or sliding bed of concentrated
solids settled at the bottom of
a horizontal pipe; and, even the
most talented rheologist scratches
his head.
The state of the art, at present,
is that calculations and predic-
tions can be made in two-phase
flow with doubtful accuracy, if
you can be sure that all of your
particles are spheres of the same
size and density. A lot of work
some of it is flowing at all! The in a vertical pipe). They also mi- by very talented scientists con-
largest particle reported to the grate and concentrate around the tinues to go on; and progress is
author as having flowed through walls of the pipe. When the being made. But look at all of the
a 6 inch pipe line was a 12 foot particles tend to move down- variables in sewage sludge. The
extension ladder. The mainten- stream slower than the carrier, particles aren’t spheres - - or
ance superintendent who report- they migrate to the center of the cubes or discs, but each with a
ed it stated that it had to have pipe. Buoyant particles with the different shape, a different drag
been pumped through the line same density as the carrier flow coefficient, a different surface
since that was the only way it in a ring surrounding the axis, to volume ratio, etc. The particle
could have gone into his new di- about half way between the cen- densities vary from lighter than
gestor. The contractor swore he ter and the wall of the pipe. A water through hard heavies. The
didn’t leave it; so, there was no change in the amount of fine par- particle sizes vary from fine mi-
other way to explain it’s presence. ticles in suspension changes the cron silts to 12 ft. extension lad-
The only mention here of the fall-out velocity of a coarse par- ders. And just when you’ve had
theory of heterogeneous flow is ticle. A small percentage of fine them all completely classified for
to say, it’s obvious that the particles in water actually re- your particular sludge and begin
total energy gradient or head duces the friction head loss. Tur- your calculations, you find that
loss for a two phase water slurry bid water flows at a greater vol- what was a particle yesterday, is
would be equal to the head loss ume under the same head than a gas bubble today, or has con-
for the water alone, plus energy clear water. Apparently, in a so- glomerated into a different shape
loss as a result of suspended lution with a very small per- and density completely.
solids. How to calculate this centage of flowing solids the par- You can easily see that loca-
energy loss is where the theorists ticles act as guide vanes and re- tion of the plant, type of grit re-
differ. The number of variables duce the water turbulence. When moval, type of sludge, mechani-
involved are so many that the cal action from preceeding op-
task of purely theoretical analysis FIGURE 6
is hopelessly impossible at the
present time. Scientists are still
working with graded spherical
particles, etc, trying not to find
out why the solid particles are
behaving in a particular manner;
but, what is their particular man-
ner of behavior. They’re a long
way from predicting sludge be-
havior with high accuracy; but,
some of the things that they do
find are interesting and applic-
able. Particles of reasonable den-
sity, under 300 microns, tend to
flow in suspension while larger
particles tend to move down-
stream faster than the fluid (such
as heavy particles flowing down
Page 5
FIGURE 7

eration, chemical treatment and conditions. This must be with ft/sec velocity. Using our com-
many other variables will change the full realization that the flow puter and data reported in as
the complete flow characteristics characteristics of no two sludges many papers on sludge handling
of the sludge. Anyone who says will he exactly the same; and, as could be found, one of our
he can accurately predict the ac- a sizeable margin of error exists. engineers, Harry Conrad, worked
tion of any sludge in a pipe line Speaking in generalities, sludges up a series of sludge friction
is misleading you and himself. below 3% are usually loss tables or curves for 4”, 6”,
What can be done then? Ob- (as mentioned previously) near and 8” pipe (figure 7), plotted
viously some estimate has to be Newtonian; and, multipliers to by percent of solids for sludge
made. Even laboratory tests tend water curves for friction loss are bands varying from raw through
toward irrelevance. More than valid regardless of the shear complete digestion.
one engineer has haphazardly rate or velocity. An increase of
opened a bulged plastic bottle approximately 25% of the value CONCLUSION
of sludge shipped in from Texas, for water should suffice. When
or where-ever, and discovered the percent of solids is higher The wide range of possible
that the properties have changed than that, the thixotropic or pseu- solid-liquid mixture makes it very
from the time when it was pour- doplastic properties begin to take unlikely that information about
ed into the bottle to the time over; and, a constant multiplier the friction loss of a particular
when he wiped it from his face, to water curves is no longer mixture will be listed in some
clothing, lab walls, . . . . Ob- valid. One test indicated 9% book or table. But consolidation
viously whatever method is used sludge to have friction losses of of available data complemented
to calculate sludge pipe friction almost 10 times that of water by good judgement and a reason-
losses, it must be empirical and of 1 ft/sec pipe velocity, while able safety factor should enable
based on past experience with the same sludge had only 3 times any competent engineer to ob-
similar sludges at similar flow the friction loss of water at 3 tain reliable estimates.