Proceedings of the Institution of Mechanical

Engineers
http://pme.sagepub.com/

Hydraulic Transport of Solid Material in Pipes

R. C. Worster and D. F. Denny
Proceedings of the Institution of Mechanical Engineers 1955 169: 563
DOI: 10.1243/PIME_PROC_1955_169_064_02

The online version of this article can be found at:

http://pme.sagepub.com/content/169/1/563

Published by:

http://www.sagepublications.com

On behalf of:

Institution of Mechanical Engineers

Additional services and information for Proceedings of the Institution of Mechanical Engineers can be found at:

Email Alerts: http://pme.sagepub.com/cgi/alerts

Subscriptions: http://pme.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://pme.sagepub.com/content/169/1/563.refs.html

>> Version of Record - Jun 1, 1955

What is This?

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 563

Hydraulic Transport of Solid Material in Pipes

By R. C. Worster, B.Sc. (Eng.) (Graduate)*, and D. F. Denny, B.Sc. (Eng.), Ph.D.
(Associate Member)t
The growing interest in the hydraulic conveying of solid material in pipes has brought to light the
lack of data for the design of these systems. To meet this need extensive research work has been
undertaken both in Britain and overseas and the more important results are summarized.
The flow of mixtures in horizontal pipes is a complex phenomenon not very amenable to theoretical
study. It has been found possible to correlate data obtained over a fairly wide range of conditions
and the flow in vertical and sloping pipes is also mentioned.
The passage of solids through machines and pipes leads to breakage of the particles and wear of
the metal parts. Some test results and conjectures on this are given.
Because high pressures are required for long-distance hydraulic transport or for vertical lifting
through pipes, the practical problems of pumping solids are severe. These are discussed with special
reference to coal-feeding devices.

INTRODUCTION Notation.
For many years the hydraulic transport of solids in pipes has C Concentration by true volume in the mixture delivered.
been successfully applied in a few specialized fields such as CD Drag coefficient.
dredging, land reclamation, and the winning of some minerals. D Diameter of pipe.
One important installation is that at Copper Cliff, Ontario, d Solid particle diameter.
where 4 million tons a year of tailings are transported by pipeline
Ordinary water friction coefficient in hf = f jj, g
L 02
over a distance of 73 miles. Another pipeline 53 miles long is f
being used to transport 1 million tons a year of fine coal to a z Hydraulic gradient along pipe carrying solids, ft. of water
power station at Carling in France. This pipe is 15 inches in head per ft. of pipe.
diameter and is fed by five 400-h.p. centrifugal pumps. Even for Hydraulic gradient along pipe with water alone at same
comparatively short distances, however, the power requirements speed.
are large, especially when big solids are to be moved: as an 4 Volumetric concentration in pipe.
example, in the construction of the Beauharnois Power Canal S Specific gravity of solid particles referred to the density
in Canada one 1,500-h.p., %inch pump was used to deliver of water.
120 tons per hr. of rock crushed to 5 inches, through 2,000 Nominal velocity in pipe, ft. per sec.
feet. Net velocity of solid particles, ft. per sec.
Existing installations for hydraulic transport utilize special Free falling velocity of solid particles, ft. per sec.
centrifugal pumps whose impellers are capable of passing the Real speed of water in vertical flow.
largest pieces likely to occur in the mixture. Sometimes the , I
Density of solid particles.
solids and water in the desired proportions are fed to a hopper yL Densky of liquid.
connected to the suction branch of the pump. The pump then
delivers the mixture into the pipeline at a pressure sufficient to
TJ
?;# Overall efficiencjr of pbmp.
--
Overall efficiencv of hvdraulic liftinz svstems.
overcome the friction resistance and any static head. Where the
solids are to be recovered, screening or other separating plant is
used at the delivery of the pipeline. In suction dredging, prior E X P E R I M E N T A L TEST P L A N T S
mixing before the pump is not possible and it is necessary to rely Several experimental installations have been used to study the
on the skill of the operator to keep the suction mouth-piece close flow of mixtures in pipes. In the 3-inch pipe system, the mixture
enough to the river bed to entrain a large quantity of material was kept in suspension by agitation in a 700-gallon storage tank.
without starving the pump of water. A centrifugal pump, driven by a 25-h.p. variable-speed motor,
There is considerable interest today in the possibilities of was fed from this tank and the mixture delivered along the test
hydraulic transport in other fields, such as the moving of coal pipeline and returned to the storage tank. For flow measurement
in bulk. Earth- and debris-moving plants are often required on the delivery was diverted into a weighing tank. Part of this
temporary sites and the simplicity of hydraulic transport is then installation is shown in Fig. 1, Plate 1. With this test plant the
the major factor in determining its widespread use. For per- secondary methods of measuring flow and concentration, used in
manent installations, however, such as the feeding of coal to a a later 6-inch pipe system, were tested.
power station, reliability and economic factors are relatively In the laboratory, it is advantageous to be able to work with
more important. For this reason more precise knowledge of the high concentrationsof solids and at low velocities in the unstable-
mechanism of the flow in pipes is needed so that reliable flow region. It has been found that experimental systems are
estimates can be made of the power consumption and the size much more stable when a feeder is used to introduce the solids
and cost of the plant required for major projects. Large-scale into the pipeline after the water has been pumped up to pressure
laboratory and plant studies are now in progress to provide the separately. With such an arrangement it is easy to deliver, at
basic information on plant operation and the design of new low speeds, mixtures of coal and water at concentrations up to
machinery. 35 per cent by true volume through pipe systems similar to that
shown in Fig. lb, Plate 1.
The MS. of this paper was received at the Institution on 15th A small installation using 13-inch transparent pipe has been
October 1954. For a report of the meeting in London, on 25th March very useful. In this plant the solids are injected by the original
1955, at which this paper was presented, see p. 586.
1* Deputy Director of Research, British Hydromechanics Research I t iWhen
n c h reciprocating feeder made in 1952.
the flow of mixtures in pipelines is being studied the
Association, Harfow, Essex.
t Research Engineer, British Hydromechanics Research Association, diaculties of measurement are considerably greater than with
Harlow, Essex. water alone. Besides the flow rate, there is also the concentration

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

564 HYDRAULIC TRANSPORT O F S O L I D MATERIAL I N P I P E S
to be controlled and measured, and it is difficult to know whether liquid (Fig. 2 , Plate 1). Re-mixing of these two layers may be
steady conditions obtain throughout the system. Test results are achieved by cross-turbulence in the liquid but this occurs only
often not repeatable to better than f10 per cent. at very high speeds or with relatively fine particles (less than
Q inch).
I n a vertical pipe, the flow conditions are simpler. Gravity
P L O W O F SOLID-LIQUID MIXTURES IN PIPES forces on the particles act in the same line as the pressure gradient
The behaviour of mixtures of solid particles and a liquid and average velocity : the separating effect is then easily over-
flowing in a pipe differs from that of a simple liquid, especially come and the solids are uniformly distributed across the pipe
when the particles are denser than the liquid and their size is section.
greater than about inch.
In a horizontal pipe, gravity tends to stratify the mixture into Flow in Horizontal Pipes. The variation of pressure drop
two distinct layers : a lower layer of solids and an upper one of along the pipe with the mean velocity is shown in Fig. 3 for the
flow of coal-water mixtures along 3-inch and 6-inch pipes.
These results are from laboratory tests using coal whose maxi-
mum size was about one-quarter of the pipe bore in each instance.
The actual delivered concentration at each test point is marked
and the curves for constant concentration are interpolations. No
method has yet been discovered of feeding a closed-loop labora-
tory pipeline so that the concentration of solids is kept constant
over a large range of velocities.
At very high speeds, the pressure gradient along the pipe is
little more than that of water alone, flowing at the same mean
speed. If account is taken of the higher density of the delivered
mixture the difference is still less, but, contrary to earlier
suggestions (Durand and Condolios 1952*; Worster 1952;
O’Brien and Folsom 1937), does not seem to vanish completely
with solids above about 4 inch. In Fig. 4 the data from the

VELOCITY- FT. PER SEC.

a &-inchcoal; 3-inch pipe.

a
w
0.10
0.09
0-08

0.07
+-
:ONCENTRATION--
TRUE VOLUME,
Fig. 4. Curve of fi-iw)/c Plotted Against Velocity
$-inch coal flowing with water in a 3-inch pipe; the excess pressure
drop is proportional to concentration.

3-inch pipe are shown as the observed difference between the

E 0.06 pressure gradients (measured in ft. of water head per ft. of pipe)
5 with and without the coal; it seems fairly certain that the
g 0.05 difference does not fall to zero but to a constant value of O.lc,
a
rT
ui
but for practical purposes the increase in pressure gradient due
to the solids at high speeds is relatively small.
3 0’04 As the speed diminishes, the pressure gradient for the mixture
departs progressively from that for the liquid alone, until it
5
5 0.03
reaches a stage where it is not greatly dependent on speed but
-
u only on the concentration and specific gravity of the solids. Still
a further reduction in speed seems to cause a rise in the pressure
&
0 gradient. It is, however, difficult to obtain data in this region,
u where the pressure gradient (with a fixed solids concentration)
3 0.02 is rising with diminishing speed, because the system is unstable
e
d
(section on Stability and Safety below).
Correlation of data on flow of mixtures of large solids and
water. It will be noticed in Fig. 3 that the curves for a given
concentration of coal are similar : they deviate from the clear-
water curve at about the same pressure gradient, then fall to
about the same minimum and rise again. Although they are
0.01 similar to one another in regard to pressure gradient, they differ
OClTY
4 5
- FT. PER SEC
6 7 8 9 from one pipe to another i n the velocity at which these changes
take place. This suggests that similar velocities in pipes of
b 1)-inch coal; 6-inch pipe. different sizes should occur at equal pressure gradients : that is
Fig. 3. Pressure Gradients Along Pipes Carrying when i, or v2/gDare the same.
Coal-Water Mixtures Since it is possible to plot all the data for a particular solid
material and pipe on a single curve of (i-iw)/c against v*, as in
Actual concentrations in percentages are shown beside each test
point. * An alphabetical list of references is given in an Appendix.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 H Y D R A U L I C TRANSPORT OF S O L I D MATERIAL I N P I P E S 565
Fig. 4, and since the velocity scale of different pipes should be large solids in water alone. No quantitative information yet
as the roots of their diameters, data with different pipe sizes exists on this aspect.
should agree when (i-iw)/c is plotted against v2/D. Effect of solid and liquid densities. In some circumstances it
T o account for different densities of the solid material, it is may be practical to use a liquid of density greater than that of
necessary to divide both ordinate and abscissa by the submerged water. With a simple liquid such as brine, laboratory tests have
specific gravity of the solids (s- l), that is, to plot (i-iw)/c(s- 1) shown that marked reductions in velocities and pressure gradients
against zG‘/gD(s-1). The acceleration of gravity has been are possible. They still conform to the correlation shown in
introduced to make the abscissa dimensionless and to write it Fig. 5, if the pressure gradients are expressed in ft. of liquid
in the conventional form of a Froude number. Instead of v2/gD, head per ft. of pipe and the solids density correction (s-1) is
the hydraulic gradient of the water flow may be used, since referred to the liquid density, that is, (s-1) = (yS-yL)/yL.
i, = fv2/2gD, where f is the ordinary friction coefficient which Although the use of brine might be economically attractive for
does not vary very greatly. This last method of plotting pressure- transporting coal (it would reduce (s-1) from 0.4 to about 0.2
drop results is used in Fig. 5, which includes test data on the thus allowing a 1.4/1 reduction in velocity and a 2/1.2 reduction
in pressure gradient), it is not desirable because of the con-
tamination of the delivered coal. The suspension type of heavy
medium is commonly used in washeries today and the technique
of recovery of the medium exists. There seems no reason why
advantage cannot be taken of this kind of heavy medium in any
proposal for transporting large coal over long distances, with
considerable reductions in costs.
Solids of non-uniform densities have not been investigated
properly. It has been observed that, with a mixture of gravel
and coal, the gravel tends to collect together and travel in ‘dunes’
at the bottom of the pipe. In some circumstances this may be
dangerous but the overall pressure drop along the pipe is not
likely to be much greater than that for the mean density of the
solids.
Effect of the pipe surface. When the physical properties of the
pipe wall differ from those of steel, or when the solids trans-
ported differ from such materials as gravel or coal, significantly
different pressure drops may be observed. For instance, lining
&/(s-l) a pipe with polytetrafluorethylene (P.T.F.E.) reduces the
Fig. 5. Curve of (i-im)/c(s- 1) Plotted Against i&- 1) excess pressure by about 25 per cent. With rubber particles
3-inch pipe; &-inchcoal. small in comparison with the pipe, a reduction of pressure drop
A 4-inch pipe; &-inch gravel. below the normal also takes place (Durand 1953): however,
-t- 6-inch pipe; I-inch coal. large rubber particles tend to bridge across the pipe section
n 6-inch pipe; &-inch gravel. with consequent increase in the pressure drop. The roughness
of the pipe walls and aging effects will not usually be important
pressure drops in 3-, 4-, and 6-inch pipes, passing both gravel- because the presence of hard materials like sand or coal will
water and coal-water mixtures. smooth the pipe wall.
No satisfactory theory of the motion of mixtures of solids and
liquids in horizontal pipes has yet been advanced and the Flow in Vertical Pipes. I n upward flow the solid particles
correlation shown in Fig. 5 for the category of large solids is slip backwards at about their free settling velocity, vs (Fig. 6).
merely an empirical deduction from test data. Similar correla- The water consequently travels rather more quickly than the
tions have been published by the present author (Worster mean speed, and the solid particles rather more slowly, the
1952) and by others (Durand and Condolios 1952), but they are difference being dependent on the ratio of the slip velocity, v,,
all subject to the same qualification.
+
Effect of particle size. Above about inch or & inch, increas-
ing the size of the solids appears to have little direct effect on the
pressure drop. However, if solids below this size are to be trans-
ported, considerable diminutions in the pressure drop and
permissible velocity will be found until, with particles of the
order of 40 microns or thereabouts, no significant differences
between the mixture flow and the flow of a liquid of the same
density may be expected so long as the flow is turbulent. With
solid particles between this minimum size and & inch the
turbulence in the liquid is capable of holding particles in
suspension, and these will then contribute very little to the
pressure drop along the pipe. The proportion of particles in
suspension naturally increases with velocity or with diminishing
particle size or density. This type of flow has been dealt with in
an earlier publication (Worster 1952), and Durand and Condolios
(1952) have also taken account of this size effect in their correla-
tion.
Little is yet known about the effect of mixed sizes of solids.
When they are all above about 7; inch, experiments demonstrate
that there is little or no effect due to size grading so long as the
particles are all greater than Q inch. When the solids are made up
of a mixture of very large and very small particles it is not
possible to say what will take place. At low speeds the large solids
travel as a deep sliding bed and a great deal of the smaller
material is trapped in the interstices of the larger. At high speeds
the fine material may be completely suspended in the liquid, thus
making a ‘heavy medium’ in which the large solids are carried
along as if they were in a liquid of greater density; that is, with a
pressure gradient less than that corresponding to that of the Fig. 6. Vertical Flow Upwards

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

566 HYDRAULIC TRANSPORT O F SOLID MATERIAL I N PIPES
to the mean velocity, v, and also on the concentration of solids centration actually present in the pipe is only increased to 20.3
being transported. Simple arithmetic leads to the relation shown per cent and the real water speed to 6.1 ft. per sec. These two
in Fig. 7 between the real water speed, vw, and the nominal slight increases correspond to a change in the pressure gradient
speed. shown by a water manometer from 0.120 to 0-128 ft. per ft. of
In upward flow in a vertical pipe the volumetric concentra- pipe. In view of the difficulties of measuring concentration, flow
tion in the pipe, q, is slightly higher than the delivered con- and pressure drop to this degree of accuracy, it is not surprising
centration, c: the ratio q [ c also depends on vs[v and on the that this small increase has not been distinguished in laboratory
tests. The pressure gradient in excess of the static lift in a
vertical pipe can therefore be taken to be i = i,+c(s-l).
Fiow in Sloping Pipes. Tests have been made with a 1)-inch
pipe system on the pressure drops in pipes inclined both up and
down, at several angles. They served to confirm that the excess
pressure drop caused by the solids in the sloping pipe was equal
to the sum of the excess pressure drops in the corresponding
horizontal and vertical pipes joining the same end points, as
shown in Fig. 8. This result has not been checked for larger

1.0 2-0 I
z1 /ns

4 On concentration, q, in pipe with vertical flow.

1.6
I .5 . t.m t d ih
I=
I.4
Fig. 8. Pressure Drops in Sloping Pipes
I.3

I .2 pipes. As in vertical pipes, the static lift in feet of water head

should be taken into account in order to obtain the difference
5 1.1 in pressures as measured by gauges at the bottom and the top
of the pipeline.
I .o
Stability and Safety. Although it is desirable from the aspect
of power economy to work near the minimum in the pressure-drop
0.9 curves (Fig. 3), the system is not very stable when the ordinary
flat characteristic of a centrifugal pump is applicable: a slight
0.8 drop in pump speed or increase in the solids feed will make it
.O impossible for the two characteristics to meet and the flow will
v /us diminish to zero, while at the same time the pipeline will fill
with solids. It is therefore necessary to work to the right (Fig. 3)
b On real speed, vW,in vertical flow. of this minimum by an amount which depends on the speed
variation to be expected of the pump and the expected variation
Fig. 7. Effect of Particle Slip in solids feed. In a long pipeline the friction Characteristiccannot
change suddenly but only in a time of the same order as the
concentration delivered, and is shown in Fig. 7b. The weight ‘transit time’ of the solids through the pipeline. There is then
of the solids in the vertical pipe is taken by an additional the possibility of measuring the pressure drop along about the
pressure gradient amounting to q(s-1). The part of q(s-1) first 20 feet of the pipe near the feeding end, and of using this
corresponding to the delivered concentration, that is, c(s - l), to control either the solids feed rate or the pump speed. The con-
is not wasted energy but useful work done in raising the solids, centration present in this pipe length would be an equally useful
and only the remainder &(q--c)(s-l) is actually wasted in control signal.
the slipping of the solids.relative to the water. (This term is It can be a great help if the pump characteristic is steep near
positive for upward flow when q >c, and negative for downward its working point. As Durand and Condolios (1952) have pointed
flow when q<c.) out, a pump driven by an engine can be made more stable
The pressure gradient due to the friction of water on the pipe by locking its governor : if the pipe resistance increases the flow
walls will be of the same order as that for water alone, will diminish and the power taken by the pump will fall. The
i = f(v*/2gD), but since the actual water speed slightly exceeds engine would then speed up, increasing the pump head which
v owing to the slipping of the solids, the water-friction gradient may then be sufficient to overcome the increased resistance. This
ought to be increased slightly. There are therefore two apparent ‘steepening’ of the pump characteristic can be estimated from
sources of additional energy loss in vertical upward flow, the the head and torque characteristics of the pump and the torque
excess density gradient (q-c)(s-1) and the slight extra water characteristics of the engine. A commutator motor can give a
friction. These two terms are, however, negligible at all ordinary similar drive characteristic.
pipe speeds. For instance, with 20 per cent of $-inch coal The stability of the system obviously depends on whether the
flowing in water up a 6-inch pipe at a nominal speed of 6 ft. per solids feed is a constant mixture concentration or a constant
+
sec., the slip velocity of the coal is about ft. per sec. ;the con- mass flow of solids. One curve for the latter is shown dotted in

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 HYDRAULIC TRANSPORT OF S O L I D MATERIAL I N P I P E S 567
Fig. 9a and it is clear that this condition is less stable than a In an emergency it is possible for tightly packed solids to move
constant-concentration feed to the pipe system. along a parallel straight pipe, but in flowing around a pipe bend
The stability of flow in a vertical pipe is not a serious problem the packed bed must shear internally. In doing this there is a
as may be seen in Fig. 9b. Naturally, it is desirable to empty the dilatation and the solids are very likely to seize and stop the
system of solids before the flow is stopped, and provision should flow. It is usually at a bend or restriction that a pipeline carrying
be made for doing this in an emergency.
TABLE1. MINIMUMSPEEDS FOR REASONABLE
CONCENTRATIONS OF COAL OR GRAVEL

Pipe diameter, Minimum speed, ft. per sec.

inches
Gravel 1 Coal
1 3 13
7
3
6
9
10
13
266
12 15 76
18 17% 83

solids will choke and it may be desirable to provide means of

emptying the pipe at such danger spots.
Settling Velocity of Particles. The velocity at which a solid
particle falls in a liquid is an important factor and, in many ways,
is a good measure of the hydraulic effects of the size, shape, and
0 ;
3
j
4
I
5 6
I I
7
’ 9
I
10
I
II
I
12
~

13
;
14
I
15
116
density of the particle. For spheres it is easy to calculate the
settling velocity by equating the fluid drag to its submerged
VELOCITY-FT. PER SEC.
weight :
Q For flow in horizontal pipe.
046
0.14
where the drag coefficient cDis a known function of the Reynolds
number vsd/v.Graphs showing this function have been published
w
g 0.12 frequently, for example, by Durand and Condolios (1952) and
Worster (1952).
!
d
i
0.10 When the particles are not spherical but of irregular shape
Y their settling velocity is reduced, sometimes greatly. Particles of
a.
&
0.08 ordinary irregularity such as gravel or coal have a settling velocity
3
UJ

0.06
of about a half or a third that of a sphere of the same ‘sieve size’.
The presence of a pipe wall diminishes the settling velocity
c’ 0.04
Y
below that of a freely falling particle. A single particle falling in
a pipe three times its diameter has its settling velocity halved.
0.02 However, this effect does not seem to be present with swarms
of particles falling: tests on the effect of concentration on the
0 settling velocity of large particles indicate a linear reduction in
0 I 2 3 4 5 6 7 8 9 1011
VELOCITY-FT. PER SEC.
settling velocity with increasing concentration: at 30 per cent
concentration by volume the settling velocity has fallen to about
b For flow in vertical pipe. 0.6 or 0.7 of the free falling velocity of single particles. This
reduction is greater with finer solids.
Fig. 9. System Characteristics In most instances the settling velocity of solids in water is
small, as may be seen in Table 2.
The maximum size of solids that can be passed through a pipe
is not easy to fix. At low concentrations and high velocities single
lumps almost as big as the pipe will even pass safely around pipe
bends. It seems that the essential requirement at moderate speeds
is that the sliding bed of solids should leave sufficient free space Screen Free fall, ft. per sec. Settling velocity at
above it for the largest particles to turn over freely. Clearly, size, 30 per cent concentration,
inches ft. per sec.
therefore, the maximum size of solids being transported depends,
at these speeds, on the concentration in the pipe. For instance,
in a 6-inch horizontal pipe a rather low speed for handling coal
1 Gravel 1 Coal I Gravel I ~

Coal
is about 5 ft. per sec. At this speed the concentration present in
the pipe is about 28 per cent when the delivered concentration
is 20 per cent. Since the voids of a closely packed bed of coal are
about 50 per cent, the coal bed occupies a depth of slightly more
than half the pipe diameter, leaving only 3 inches of clear space
above it. In this case the maximum safe size of coal would be
about 2 inches. At higher concentrations or lower speeds this As may be seen from Fig. 7 a a mean velocity of double the
would be reduced. The rule that the maximum size of solids is settling velocity would be quite adequate for transport up a
limited to one-third of the pipe diameter is only applicable at vertical pipe, that is, 3 ft. per sec. for 3-inch coal. But if a coal-
moderate concentrations if the speeds are low. raising system is made of the same pipe throughout then the
The speeds that correspond to the minimum in the friction velocities must be determined from the requirements of any
curves increase as the square root of pipe diameter and of the horizontal or sloping sections in the pipeline. These velocities
density under water (s-I). Approximate values are given in will always be large in comparison with those sufficient for safe
Table 1 for reasonable concentrations of 20 or 25 per cent of vertical lifting, but it hardlv seems worth increasinp: the size of
-
I

large coal or gravel. 1 the vertical $pe on this account.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 HYDRAULIC TRANSPORT O F SOLID MATERIAL I N P I P E S Plate I
Mean speed
of mixture,
ft. per sec.
5.5

4.5

3,7

2.8

1.3

0.8

a Fig, 2. Flow of a Coal and Water Mixture at Different

Speeds in a Perspex Pipe of l+ inches Diameter
Delivered concentration about 13 per cent.

b
Fig. 1. Experimental 3-inch Pipe System at the British Hydromechanics Research Association (B.H.R.A.)
[I.Mech.E., 19551
Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014
Plate 2 HYDRAULIC TRANSPORT OF SOLID MATERIAL I N PIPES

Fig. 18. Part of Ring Pipe Used for Measuring Coal Degradation Fig. 20. Wear on the 8-inch Hammersmith Coal Pipeline
Wall thicknesses of 0.25 inch and 0.15 inch at top and bottom
respectively.

SCREEN

SURFACE
PLANT CIRCULATING PUMP
w
SUMP
i

PRESSURE VESSELS

RETURN
WATER

DISCHARGE VESSEL

1 , WATER DISPLACED
B Y COAL

:
i:l
ANNULAR

PIT
BOTTOM

FEEDER Fig. 12. Lock-hopper Feeder IET

Fig. 10. Typical Closed-cycle System Employing a Feeder

[I.Mech.E., 19551

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

568 HYDRAULIC TRANSPORT O F SOLID MATERIAL I N P I P E S
For instance, if the horizontal pipes are chosen to be of 10- avoided. However, the fines that would otherwise accumulate in
inch bore with a mean speed of 9 ft. per sec., the water friction the system must be withdrawn at a rate sufficient to keep their
losses in the vertical pipe would be about 0.03 ft. of water per concentration at an acceptable level.
ft. of pipe. A change to 12-inch pipe for the vertical line only A closed cycle would be essential for lifting coal or other
would reduce the speed to about 6.25 ft. per sec. and the pres- mineral from a dry mine, and for systems using liquid media
sure gradient to 0.012 ft. of water per ft. of pipe. In a 1,000-foot other than water.
lift this represents a saving of perhaps 12 feet of water friction,
which is hardly significant. Practice may show that other factors Eficiency of a Hydraulic Lifting System. With a vertical
such as pipe wear make the use of lower vertical-pipe velocities lifting system where work is done on the solids it is possible to
desirable. arrive at a conveying efficiency, and a typical closed-cycle
installation is analysed below.
M E T H O D S OF P U M P I N G S O L I D S The energy is provided partly by a circulating-water pump
Low-pressure Systems. The simplest method of pumping used to overcome friction losses and the density difference of
solids is to mix them with the desired amount of water in a tank the two water columns, and partly by a pump at the pit bottom
and then pass them through a centrifugal pump the passages of which returns the displaced water and feeder leakage back into
which are wide enough to take the largest particles. If the pump the system at the highest pressure point. From the practical
speed and solids feed rate are constant there is no difficulty in point of view, the obvious location for the circulating pump is at
maintaining a uniform flow in the pipe. the surface where the pressures are low.
This method is widely used for conveying fine material, and The feeder does no useful work: it simply exchanges equal
one notable scheme is that at Carling, where five centrifugal volumes of solids at low pressure for water at high pressure.
pumps in series convey colliery tailings to a power station over a The chief power loss in the system arises from the leakage of
distance of 53 miles (Gibrat and Chenin 1954). high-pressure water to atmosphere through the feeder.
It is usual with solids pumps to pressure-grease the internal The effects of feeder leakage and of solids concentration on
bearings and to flush the glands continuously with clean water, system efficiency are shown in Fig. 11, an 80 per cent efficiency
but the real problem is the erosion of the volute and impeller
and it is this factor which restricts their use. For instance, when 1.0
handling solids the pump speed is limited and the maximum
head available is about 150 ft. per stage, so that a single pump is
sufficient only for comparatively short-distance transport unless
0.8
staging is employed. Even with fine coal, erosion of high-speed
pumps may be serious.
High-pressure Systems. Long-distance horizontal transport 0.6
or vertical lifts from deep mines necessitate a very high pressure c
F
at the entry to the pipe system, or boosters along its length. This
can be achieved by any of the following methods : > 0.4
(1) The use of several centrifugal pumps in series.
(2) The solids and water may be passed together through a
positive-displacement pump. The problem of valves for such a
pump is a serious one, for their lift or opening must be sufficient a:
to allow the largest particles to pass freely through, and they are
also subject to damage by the particles on the sealing surfaces.
So far as is known this method has not been used for solids
+
larger than inch, but there is the possibility of a special valve 0
r b h - 0 20
for a pump being designed to handle much larger solids than this SOLIDS CONCENTRATION-PER CENT
with an efficiency better than that of the corresponding centri-
fugal pump. Fig. 11. Effect of Solids Concentration and Leakage on
( 3 ) The solids may be fed mechanically into the pipeline after Lifting Efficiency, 71, of a Closed-cycle Feeder System
the water has been raised to the necessary pressure. Several Feeder power is neglected ;KLequals (feeder leakage)/(solids through-
designs of feeding machine have been tested and these will later put); rlp is the displaced-water pump efficiency.
be described more fully. It is the essence of all such machines
that they release water to atmospheric pressure at the same being assumed for the circulating pump. It can be seen that the
volume rate as the solids feed, and this, together with leakage efficiency of the cycle is greatly affected by the solids con-
water, must be pumped back into the pipe. The pressure of the centration when below 15 per cent, because power is required to
system at this point is at its highest and the repumping of the maintain the water circulation, and the friction component of
displaced water which may be heavily laden with fine material this power is independent of the solids delivery.
presents a problem. Design of suitable centrifugal pumps for These calculations neglect the power to work the feeder which
this duty may be difficult and, in many instances, the size and is zero only in the ideal case. Power is required to overcome
cost of cleaning plant for this purpose would be prohibitive. A mechanical friction and, where used, to suck the solids into the
positive-displacement pump may often be a better solution. feeder from the hopper. Both terms will depend largely on the
feeder design but may be expressed as a function of the solids
System CjJcZes. Hydraulic transport systems may be divided throughput.
into two categories, closed cycle and open cycle, according to If the feeder power consumption is 0.3 h.p. per ton per hr. of
whether the liquid conveying the solids is recirculated and used solids, with a 1,000-foot lift, 50 per cent leakage, and 20 per cent
again or not. concentration, T / Q falls from 0.72 to 0.63 (if ~p equals 0.75)
Open cycles are applicable only where there is ample supply of when feeder power is taken into account.
liquid near the inlet end of the pipeline, and means of disposing Thus, the lifting-system efficiency would be about 48 per cent.
of the liquid at the discharge end of the pipe. The conveying
liquid will be contaminated by fine particles and, in an open-cycle Types of Solids-feeding Machine. Feeders may be divided
system, it may be difficult to dispose of, or uneconomic to clean. into two categories, those in which the only moving parts of the
A typical application of the open cycle is the lifting of coal from machine are valves, and those in which the moving parts them-
a mine where water is available at the pit bottom. selves convey the solids from a region of low pressure (hopper)
In the closed cycle (Fig. 10, Plate 2) the water used to convey to a region of high pressure (delivery line). In practice, however,
the solids is returned, after separation at the discharge end, and the problems of both types are almost identical. Except for the
used again. I n this way, both the provision of a continuous supply lock-hopper feeder, machines for feeding solids other than coal
of water and its subsequent disposal after contamination are into high-pressure pipelines have not been developed.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 H Y D R A U L I C T R A N S P O R T O F S O L I D MATERIAL I N PIPES 569
One design of the first type, a lock-hopper feeder made by necessarily become damaged, should be relied on to seal the
a commercial company for the Ministry of Fuel and Power, leakage.
consists of two pressure vessels each communicating alternately There are many forms the feeder can take; two are illustrated
with a hopper and a discharge vessel as shown in Fig. 12, Plate 2. in Figs. 14 and 15.
As an indication of size the outer vessel is 8 feet in diameter, COAL
and 15 feet high; the main valves are elliptical and measure AND WATER
24 inches by 18 inches. The throughput of the machine will be
determined not so much by the time cycle of events as by the
rate at which coal can fall through the valves agajnst the up-
flowing displaced water. With 3-inch coal the particle velocity
through the valves is unlikely to exceed 1 ft. per sec., which
corresponds to a throughput of about 100 tons per hr.
The main point in favour of the lock-hopper feeder is that the
leakage from the pipe to the hopper can be very small. This is a
very important factor in the efficiency of the transport system.
The valves move only after the pressures on either side are
-e--
balanced, so that the power needed is small and erosion of the
surfaces is minimized. SUCTION
Very large vessels are unsuitable for high pressures, and it HIGH-PRESSURE
may be necessary to use small vessels with a correspondingly WATER
short time cycle. The control system is complex, there being four
coal valves and at least four water valves to be operated in their Fig. 14. Reciprocating Feeder (Double-acting)
correct sequence.
The feature common to feeders of the moving-pocket type
is that solids from the hopper are induced to fill a pocket within
the moving part of the machine, and this pocket then travels to
a position where water at high pressure drives the solids into the
delivery pipe. The throughput of such a machine is proportional
to the linear speed of the pocket, and to its volume, provided
that it completely fills and empties.
If the particles are merely allowed to fill the pockets by
gravity the speed and hence the throughput of the machine is
seriously restricted by the slow rate of filling and, in practice,
some inducement for the particles to move quickly is essential.
This can be achieved by circulating the solids and water through
the pocket, in one side and out the other, but this involves a
solids-handling pump and unnecessary degradation of the
particles. An alternative method is to circulate only the water, the
solids being restrained from leaving the pocket by a suitable
grating across the outlet, as shown in Fig. 13. SUCTION k I HIGH PRESSURE
WATER
AND
WATER

SECONDARY
Fig. 15. Disk Feeder
WATER FLOW

Fig. 13. Moving-pocket Feeder Reciprocating feeders. In the double-acting reciprocating

feeder (Fig. 14) a plunger containing two rectangular pockets is
reciprocated so that at each end of the stroke one pocket lies
Provided that the grating forms part of the moving member beneath a hopper and is filled with coal, whilst the other pocket
of the machine rather than part of the body, the grating is self- falls opposite the pipeline into which its load of coal is discharged.
cleaning, because flow through it is reversed when the solids are The delivery from a single machine of this design is interrupted
discharged from the pocket. With this method of filling the for a short time during each stroke because of the lap necessary
pocket, the throughput of the machine is now limited by the to restrict axial leakage from the pressure pipe to the hopper. This
rate of emptying the pocket, which is a function of the quantity interruption can be avoided either by allowing a continuous
flow in the pipeline and thus normally beyond control. bypass of water round the feeder, which lowers the delivered
The restriction of leakage from the high-pressure side to the concentration, or by intermittent bypass of water through radial
hopper is the major probIem for wherever seals may be fitted holes in the plunger. The 50-ton-per-hr. prototype machine at
they have at some time to pass over the rough edges either of the B.H.R.A. (details of which are given below) employs the latter
pocket or of the hopper opening; this virtually eliminates the use method, but the flow is not uniform and the best solution seems
of deformable materials such as rubber. to be to use two or more machines in parallel. A triple-plunger
The present solution is to rely on close clearances to restrain feeder for 100 tons per hr. on these lines is being constructed for
leakage, but though this is satisfactory when the machine is installation in a pit.
new, wear inevitably results in increased leakage. Rotary feeders. From a design standpoint, continuous motion
With these feeders the moving parts must be capable of is advantageous because the drive is so much simpler and the
chopping those particles which can neither squeeze into the inertia of the moving parts less than with the corresponding
pocket nor climb from it at the moment of closing of the gap. reciprocating machine. T o minimize inertia is important in
It is thus undesirable that these ‘chopping’ edges, which regard to jamming of the machine by metal objects that may fall

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

5 70 H Y D R A U L I C T R A N S P O R T OF S O L I D MATERIAL I N P I P E S
in the hopper. On the other hand, from a hydraulic point of view, applications the feeder will be chosen almost entirely on leakage
the pockets have less chance to fill and empty when moving considerations.
continuously than when reversing at each end of a reciprocating It is clear that there is less to be gained by making the leakage
stroke. path very long than by reducing its cross-section, and a length
The choice of arrangement of rotary feeders is wide, though in of 100 times the clearance is probably sufficient. Feeders having
practice many designs cannot be considered for high-pressure leakage paths of small cross-sectional areas are thus to be
applications on account of their inherently high leakage. preferred to those having long leakage paths of larger area.
One form of rotary feeder, in which all the pipe and hopper There are some results of leakage tests on reciprocating feeders
apertures lie in one end face is shown in Fig. 15. With duplicate but no comparative results for other designs.
hopper and discharge pipes, the machine is symmetrical about When new, the radial clearance between the mild-steel
a diameter, and with five or more pockets in the rotor the dis- plunger and cast-iron body of the 14-inch feeder (the size refers
charge is completely uniform. A small Perspex model has been to the diameter of the delivery pipe) was 0.0015 inch, and the
tested and has given promising results. The rotating disk must shortest leakage path 1 inch long. Leakage was proportional to
be loaded against the body under all conditions so as to keep the pressure and amounted to 5 gal. per min. at 400 lb. per sq. in.
clearance small, but the force must not be allowed to cause (920 feet of head), equivalent to about 20 per cent of the main
excessive wear. Control of the force under varying flow conditions flow for a pipe velocity of 6 ft. per sec.
will probably be the main design problem. After 40,000 strokes (12 hours) passing $-inch anthracite,
leakage was only very slightly increased. The machine was then
Performance of Coal Feeders. Some performance data are run for a further 12 hours on a coal-sand mixture designed to
available for the reciprocating-plunger type and the rotating- accelerate wear; during this test, leakage increased to treble the
disk type, though no feeder has yet been in sustained operation original amount and severe local wear occurred, though most of
under full working conditions. the sliding surface was undamaged.
The first feeder built by B.H.R.A. was a small machine on the After the worn plunger had been replaced by a new one made
lines of Fig. 14. The plunger was 2 inches in diameter and of brass, the feeder ran for about 500 hours on anthracite without
18 inches long, containing two pockets each of 8 cu. in. in noticeable increase in leakage and without serious wear on the
capacity; the stroke was 73 inches, and the overall length of the plunger, the only visible damage being slight rounding of the
machine 32 inches excluding the drive. The pipe circuit was chopping edges. It is noteworthy that the Perspex model ran
arranged so that the feeder could deliver coal into a 1-f-inchpipe on anthracite for 200 hours before wear was serious.
at pressures up to 400 Ib. per sq. in. (1,000 feet of head). Leakage As originally constructed the 4-inch prototype feeder had a
was restricted by a close clearance between the plunger and the hollow mild-steel plunger, 6 inches in diameter and 6 feet long,
body. Later a Perspex model of this feeder and also a model of supported in phosphor-bronze bearings. Leakage was restricted
a disk-type feeder, each supplying $-inch pipes, were con- by two brass piston rings, each 1 inch wide, having a radial
structed so that the flow of coal into and out of the pockets could clearance of 0.003 inch on the plunger, and amounted to
be studied. 12.5 gal. per min. at 1,000 feet of head (4.2 per cent of normal
As originally designed there was no provision in this feeder water throughput) when the machine was new. After running
for secondary flow to assist the coal into the pockets, and the 47 hours at a speed of 60 r.p.m., during which time the equiva-
coal delivery was less than one-quarter of the theoretical quantity lent of 900 tons of both graded and run-of-mine coal were passed
based on complete filling and emptying of the pockets. When the through the machine, the leakage had risen to 47 gal. per min.
feeder was tested under a pressure of 150 lb. per sq. in. these (16 per cent of water throughput) at which figure it remained
deliveries were halved by the leakage flow into the hopper. until the test was stopped at 100 hours.
However, when the secondary suction flow was introduced the When dismantled, the plunger was found to be virtually un-
coal delivery could be increased almost to the theoretical damaged and there was no measurable wear on the cylindrical
quantity, and was then unaffected by leakage flow. surfaces; the chopping edges on the pocket had become rounded.
The seal rings were scoured and worn to an oval shape with the
major diameter vertical as though coal had become trapped in
the clearance. At these places the clearance had increased to
0.010 inch compared with 0.005 inch on the horizontal diameter.
The bronze bearings were in perfect condition.
Power requirements of feeders, Since a feeder does no real
work, power is required only to overcome mechanical friction.

Fig. 16 shows the effect of secondary flow on coal delivery at

various speeds. In general, secondary flow needs to be of the
same order of magnitude as the main flow to ensure maximum
capacity of the feeder.
I t has been seen that the leakage through the feeder has an
important effect on the system efficiency, and for high-pressure

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 HYDRAULIC TRANSPORT OF SOLID MATERIAL I N PIPES 57 1
This is small as may be seen from Fig. 17 which shows the that although the pressure drop should be somewhat more than
steady power requirements of four feeders at B.H.R.A. In the sum of the water friction gradient, ,i and the density
practice, a good margin must be allowed to overcome temporary gradient, c(s-1), this difference is not distinguishable in
overloads, such as accompany the feed of a high proportion of practice. It is a simple matter to measure the difference in head
very fine coal. over 10 or 20 feet of vertical pipe, subtract the water friction
Where secondary flow is used to suck the coal from the hopper, corresponding to the speed in the pipe (which can be obtained
power is required to drive the pump; this will be directly pro- from the nozzle flow measurement), and obtain the concentration
portional to the theoretical coal throughput and of the order of from c = (i-iw)/(s-l).
0.02 water h.p. per ton per hr. For instance, on 10 feet of 6-inch pipe, running at a speed of
6 ft. per sec., a water friction drop of about 2 inches of water
gauge would be expected, whilst the density-head difference
MEASUREMENTS with coal at a concentration of 20 per cent would be about 9.6
I n simple hydraulic-transport schemes such as those used in inches of water gauge. Durand and Condolios (1953) have com-
gravel pits or for dredging, where the feed rate of solids is not bined a nozzle flow-meter with the vertical-flow concentration
easily controlled it is common practice to operate at high speed measurement in the following way : at the bottom of the vertical
and low concentration because of the danger of blocking the leg there is a nozzle whose contraction is chosen so that the
pipe. If suitable instruments to measure concentration and pressure recovery after it just balances the water friction in the
velocity were available they might help greatly in increasing the vertical test pipe; the pressure difference measured between
throughput of the system even though the power saving might the throat of the nozzle and the top of the vertical pipe is then
not be very important. I n a large and more steady system, such directly proportional to the concentration in the pipe.
as that envisaged for transport from a colliery to a power station,
considerable powers may be involved and the efficiency would
D E G R A D A T I O N O F COAL A N D WEAR O F P L A N T
then be important. In general, efficiency and safety go in opposite
directions, so that the lower the velocity and the higher the con- Degradation of Coal Travelling Along Pipes. Most present-
centration the greater is the need for precise knowledge of the day methods of moving coal do, to a certain extent, break it up.
conditions in the pipeline. I n hydraulic transport, the sliding of the layer of coal along the
bottom must do some damage both to the coal and to the pipe;
Measurement of Flow or Velocity. In view of the relative ease in particular, at bends or bad joints the whole layer of coal
with which solids are accelerated in water, a Venturi-type nozzle suddenly turns over. The top layer of coal moves along in a
was tried in a vertical pipe on the 3-inch coal system. Handling hopping motion at a speed which is little smaller than that of
water alone it was found to have a discharge coefficient of 0.95 the water, while the sliding layer underneath may be moving at
and, when handling coal-water mixtures at speeds down to 13ft. only a half or a third of this speed. Impacts must take place
per sec., the discharge coefficient fell to 0.93f2.6 per cent if the between coal moving at different speeds and between coal and
head differential was measured in ft. of water. If, however, the the pipe walls. These shocks not only abrade the large coal and
differential was measured in ft. of mixture density, the discharge reduce its market value but also produce very fine particles which
coefficient became 0.955 1 2 . 2 per cent. This suggests that then create an effluent problem.
mixture density should be used rather than liquid density in I n hydraulic-transport installations the coal would usually
calculating the differential on the nozzle. pass once through a pump or feeder and then travel for perhaps
One company (Anon. 1953) has used a similar device com- several miles along the pipeline. I t may be expected, therefore,
prising an eccentric nozzle in a horizontal pipe, which has been that the damage to the coal done in the pipeline would pre-
shown to have the same discharge coefficient with solids as with dominate over other damage except for very short pipe lengths.
water if the differential head is measured in feet of water head. I n laboratory installations the length of the pipe is limited to
With this arrangement there is a minimum speed (coyresponding one or two hundred feet, so that a large proportion of the damage
to the existence of an appreciable bed of solids) below which this takes place in the pump and feeder and it is impossible to separate
device must not be used. the effect due to the pipeline. Recirculating the coal round an
Other devices at present available for the measurement of experimental installation does not help in this problem because
velocity include one (based on electromagnetic induction) which of the many passes then made through the pump or feeder.
is manufactured in Holland for dredging applications, and To measure degradation experimentally, relative motion
another where Coriolis forces on a curved rotating pipe are used between coal, water, and pipe similar to that in hydraulic trans-
to measure the total mass flow. port along a horizontal pipe has been achieved with a toroidal
or ring pipe arranged at the periphery of a large wheel shown
Measurement of Concentration. Concentration can, of course, in Fig. 18, Plate 2. The ring pipe is filled with water until it is
be measured by sampling the delivery, but what is really wanted about three-quarters full with a 2-lb. or 3-lb. sample of coal in it :
is an instrument which can be fitted at any place in the pipeline when the ring is rotated the water and coal remain approximately
to indicate the concentration present there (which is greater than stationary while the pipe wall moves along. Water is carried
the concentration delivered). As far back as 1948, one company forward with the pipe near the walls, and since none is carried
had developed electrical-conductivity methods of measuring the over at the top of the wheel there is a return in the centre. When
proportion of air bubbles in a flowing air-water mixture and the this is combined with the velocity of the ring pipe, the resulting
same technique was applied to the measurement of the concentra- velocity of water and coal relative to the pipe is very similar to
tion of gravel and sand present in a horizontal pipe (Durand that in ordinary horizontal transport. Centrifugal force does not
and Condolios 1952). A special shape of electrode was found to arise as the pipe alone rotates. This ring pipe is of 4 feet diameter
make the instrument insensitive to the actual position of the and has a 6-inch square pipe section with the front face made of
solids in the pipeline. Perspex so that the behaviour of the coal may be observed. The
Very recently a commercial device working on this principle front face is fitted with two flush-fitting doors through which the
has appeared in Holland. Such methods of measurement involve samples of coal are introduced and withdrawn.
difficulties due to the varying conductivity of the water itself. Several series of tests on degradation of 1-inch coal have been
Where the solids have an electrical conductivity of the same made with this machine at speeds of 7 5 9 . 1 , and 11 ft. per sec.
order as the water the calibration of the device will depend on for periods corresponding to distances up to 70 miles. The coal
the ratio of the two conductivities so that it is difficult to apply samples to be tested were put in the machine, run for a specified
the method to the measurement of coal concentration. The fine- distance, and then removed: after drying, they were screened
coal pipeline at Carling is fitted with a gamma-ray absorption and weighed. A particular set of results is shown in Fig. 19, from
type of instrument which measures the total mass within the which it willbe seen that the greater part of the attrition products
pipe from which the mass concentration present in the pipe are well below 4inch. The degradation of this particular coal is
can be deduced directly (Gibrat and Chenin 1954). This again tolerable-about 1 per cent loss of weight in the first mile
is a rather complicated installation. followed by a gradually diminishing rate of loss with increasing
In the section above on Flow in Vertical Pipes it was noticed distance travelled.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

572 HYDRAULIC TRANSPORT OF S O L I D MATERIAL I N P I P E S
If degradation is defined as the attrition products below 8 inch tion loss of about 4 per cent. Great differences between one pump
it has been found that, very roughly, at each speed the total loss and another may be expected : for instance, when a rubber-lined
increases in about the same proportion as the square root of the pump was used for the same duty as the iron one referred to
distance travelled. This diminishing rate is reasonable because as above, the damage to the coal was reduced to about one half.
the coal becomes more rounded there are fewer corners to be Velocities in a coal feeder are low and independent of the head
knocked off. against which it works, so that damage caused by impact will
therefore be very slight. On the other hand, there is the chopping
I I I I I associated with the valve action of the machine. It has been
estimated that in one of the reciprocating machines only one
piece of coal in a hundred is broken by chopping, so that the
damage is quite slight.
Abrasion and Wear of Plant. In some instances where hard
particles have been transported pneumatically, a very short life
of only a few days has been reported for pipe bends. I n hydraulic
gravel-handling plant (where the velocities are about double
those required for coal transport) the pipe may last a year or
I more if it is rotated periodically so that the wear is distributed
20 around its periphery. Pipe bends may last four or six months
DISTANCE TRAVELLCD-MILES
before they require patching. Centrifugal-pump impellers made
Fig. 19. Degradation of 1-inch Coal in Hydraulic Transport of abrasion-resisting steel may need replacing in a year, even
at 9.1 ft. per sec. though their speeds are generally low. A rubber lining to a pump
or pipeline completely eliminates wear but it is liable to be cut
Over a range of speed from 7.5 to 11 ft. per sec. the loss rate by large sharp flints, or pieces of metal. The use of rubber lining
increased approximately as the cube of the speed. Within this is usually restricted to gravel no larger than about 2 inch and
range of conditions and for this particular coal the degradation precautions must be taken to keep out pieces of metal. Because
seems to be given approximately by : coal is less sharp and less dense the limiting size of coal for
Total loss cc(speed)3 x (distance)+ rubber-lined pumps might be raised to about 2 inches but there
is no experience on this point since the centrifugal pumps used
This speed relation suggests that most of the damage is not done in washeries are usually of low head, and of simple iron con-
by the dragging of the bottom layer along the pipe but by struction.
impacts between 'flying' pieces of coal and perhaps also between The damage to a pipeline when transporting coal will be much
flying coal and the side walls. This is another important reason less than when transporting gravel. The coal pipeline at Hammer-
for running a coal pipeline at the lowest safe speed. smith which was in use for more than ten years appeared to have
It was observed in the ring-pipe test apparatus that when the lost 0.1 inch of its wall thickness at the bottom (Fig. 20, Plate 2),
doors in the front Perspex face of the pipe were not replaced but it is not known how much coal was transported through this
exactly flush (causing a recess about 0.015 inch deep) the damage pipe or how often it was allowed to stand empty and corrode.
to the coal was greatly accelerated. This sort of irregularity is The pipe could have been turned round into two other positions
likely to happen at each pipe joint and it may be desirable, in a so that a real working life of at least ten years may be expected
long pipeline carrying large coal, to use special pipe joints. for a coal-transport pipe. Pipe bends are likely to have a shorter
life and there is the danger that erosion at misfitting pipe joints
Degradation by Machinery. It is more difficult to estimate the may be important.
damage done by other hydraulic devices such as centrifugal Other workers (Bergeron 1952) in t h i s field, all report a rapid
pumps or feeders. Centrifugal pumps used in the laboratory to increase in erosion with speed, usually as the cube of particle
circulate coal-water mixtures around pipe systems are presumed speed. The character of the wear depends on the angle of attack,
to be responsible for the greater part of the degradation occurring and with sliding or scraping the rate of wear increases almost in
in these tests. In two or three hours' testing with a pump head proportion as the particle size. It also increases with solids con-
of the order of 30-50 feet +-inch coal may be reduced to about centration. Naturally it also depends on the curvature of the wall
$-inch together with the production of a great quantity of fines. and on the density and hardness and irregularity of the particles.
During this time the coal would have made about 200 passes With normal impact the phenomena are rather different, for
through the pump so that it may be guessed that the degradation instance, particle size is said to have little effect.
for a single pass through a centrifugal pump is of the order of
fs or f per cent for a pump working at this speed. At higher pump
heads this would be greater, since the pump speed increases as CONCLUSIONS
the square root of head. If the relation between pump speed and The outstanding characteristics of hydraulic transport systems
coal degradation is also a cubic law it would seem that a similar are the large power required and the high capacity of pipes of
pump generating a head of about 200 feet might cause a degrada- quite modest sizes. For instance, as can be seen from Fig. 21,
1.200- 600

f
I 7

1,000- 500
500
400
400

300 f 600- 300

a
200 f 400- 200

i00 200 - I00

0 - 0 0
4 6 8 10 I2 14 16 4 6 8 10 I2 14 16 4 6 8 10 12 14 16
VELOCITY-FT. PER SEC. VELOCITY-FT. PER SEC. VELOCITY-FT. PER SEC.

a Pressure drop plotted against b Horsepower per mile plotted c Coal delivery plotted against
velocity. against velocity. velocity.
Fig. 21. Performance of a 12-inch Pipeline Transporting Coal

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 HYDRAULIC TRANSPORT O F S O L I D MATERIAL I N P I P E S 573
the pressure drop along a 12-inch pipe conveying coal may ACKNOWLEDGEMENTS
exceed 300 ft. per mile but the capacity is of the order of 250- Acknowledgementsare due to the many members of the British
300 tons per hr. of 3- or 4-inch coal. Hydromechanics Research Association who have shown interest
The compactness of the conduit gives hydraulic transport in this work and offered valuable advice, to colleagues for their
great advantages for conveying solids in confined spaces or practical assistance and to the Council of that Association for
where there are obstructions making it impossible to route permission to publish the material contained in the paper.
directly. The ease of hydraulic ‘loading’ and ‘unloading’ is also
important.
The application of hydraulic transport to the lifting of coal
from the pit to the surface is now being considered for very deep
mines where the journey time of the cage and the weight of the APPENDIX
rope limit the lifting capacity of the shaft. The advantages of
hydraulic lifting, where 200 tons per hr. could be handled by a REFERENCES
12-inch pipe, may prove to outweigh the practical difficulties ANON. 1953 La Houille Blanche, vol. 8, p. 58, ‘Debitmetre 8
involved in bringing the coal-water mixture to the very high Contraction pour les Melanges d’eau et de Materiaux’.
pressures at the bottom of the pipeline. In this connexion several BERGERON,P. 1952 ‘SocikttC Hydrotechnique de France
prototype feeders which have shown promise in the laboratory Deuxikme JournCes d’Hydraulique’, June (published by
are now to be tested under site conditions. Even for shallower La Houille Blanche), ‘Factors Affecting Wear Caused by
mines hydraulic lifting should be valuable, for the output of Solids in Hydraulic Transport’.
many mines in Britain is limited by the winding capacity of the DURAND, R. 1953 La Houille Blanche, vol. 8, p. 124, ‘Flow of
shaft. This restriction on mine output could be relieved by Mixtures in Vertical Pipes, etc.’
screening the coal (below the surface) to 3 inches and pumping DURAND,R., and CONDOLIOS,E. 1952 ‘SociktC Hydro-
the smaller sizes through a pipe installed either in the shaft or technique de France. Deuxikme Journkes d’Hydraulique’,
in a special borehole. Calculations indicate that the efficiency of June (published by La Houille Blanche; also published
such a lifting system need be little less than that of the ordinary November 1952 in Proc. Colloquium on Hydraulic
winding method. Other advantages of colliery applications lie Transport by National Coal Board), ‘Study of the
in the possibility of greatly reduced health and fire hazards due Transport of Solids in Pipes’.
to dry coal dust. 1953 La Houille Blanche, vol. 8, p. 292, ‘A Device for
Permanent pipelines working with reasonable prospect of Measuring Concentration in a Mixture Flowing Along a
safety must be carefully designed and controlled to ensure Pipe’.
stability in operation. T o a large extent this depends on the GIBRAT,R., and CHENIN,F. 1954 Inst. E. Eng. paper S.1630,
combined characteristics of the pump and the pipeline and on 24th February, ‘The Transport and Utilization of Colliery
the method of control of water and solids feed. Automatic control Tailings at the Emile Huchet Power Station’.
by instruments measuring pressure, velocity or concentration O’BRIEN, M. P., and FOLSOM, R. G. 1937 University of
seems desirable for large-scale installations working continuously. California Publications in Engineering, vol. 3, p. 343,
It seems clear that with soft solids such as coal a reasonable ‘The Transportation of Sand in Pipelines’.
life of machinery and pipes can be expected, but with hard WORSTER,R. C. 1952 Proc. Colloquium on Hydraulic Trans-
materials like sand abrasion is severe and fairly frequent replace- port, November (National Coal Board, London),
ment of parts may be necessary. Hydraulic Transport of Solids’.

Discussion
Mr. G. A. WAUCHOPE (Member), in opening the discussion, said per hr. of mixture containing 50 tons of coal. The system had
that he had followed the matter of the transportation of solids been driven by a 110-h.p. motor, and the scheme had worked
with very great interest for a number of years. As the authors more or less satisfactorily for some time. It had been a very
had said, it was not a new subjecr. It was a very old one. I n large plant for the capacity in the light of the latest information.
saying that it was more than half a century old, he thought the Since that time until about 1951, as far as he was aware,
authors were underestimating its age : it was probably much practically no interest had been taken in the subject. Then,
older than that. He believed that suction dredgers had been used simultaneously, attention began to be given to it in the United
for at least seventy-five years. States, France, and Great Britain.
Suction-dredging in the past-probably what was being Coal pumping, which had been dormant, was not the only
discussed was more akin to that than any other form of the type of transportation of solids which had been used. In the
hydraulic movement of solids-had been used as a tool by the meantime, developments in ash-handling had involved a con-
civil engineer. The most important concerns had been its siderable amount of pumping of solids.
availability, the ease in getting it to the job, and its ease of use In the period 1951-52 thought began to be given to it; and
on the job. the Americans, the French, and the British had got to work;
In the past there had not been a great effort to examine care- and the authors had played a leading part in obtaining a much
fully what happened in the pump which was used in suction more exact knowledge about what happened in the pipeline
dredging, or in the pipe. It had been found satisfactory to be than had ever been known previously. What the authors were
able to move a certain quantity of material in the simplest way doing was finding out what were the correct velocities, the
possible. correct concentrations, and the correct sizes of pipe to deal most
As the authors had mentioned, nearly forty years previously- economically with transportation problems.
perhaps, rather more than that-Mr. C. C. Bell, electrical It was vital to have that information because with a pipeline
engineer to the Hammersmith Council, had been farseeing of a certain size it was most important to keep the concentration
enough to install a coal-pumping system to pump coal from the as constant as possible. I n dredging, and in suction-dredging in
wharf by the Thames to the power station. It might be of interest particular, it was extremely difficult to control the rate of con-
if he were to add one or two details to those in the paper. The centration or, rather, to control the concentrations so that the
length of the pipeline had been, he believed, 2,035 feet, and the rate of flow of solids was constant. Therefore, everything had to
pipe had been 8 inches in diameter. Two gravel pumps connected be done on a liberal scale. For example, it would perhaps be un-
in series had been employed, being supposed to pump 60,000 gal. reasonable to take as a base a higher percentage of concentration

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

574 DISCUSSION ON HYDRAULIC TRANSPORT O F SOLID MATERIAL I N PIPES
than 10 because 25-30 per cent might sometimes accidentally positive-acting pumping feeder with the bulk of water being
occur. If the thing was to be done economically the concentra- supplied from a centrifugal pump, which might be the solution.
tion had to be controlled within very fine limits. There was one minor criticism that he would like to make of
Then there was the question of the stability of the pump the paper, and that was that it laid a certain amount of stress
characteristics. The importance of that had been shown in the on experiments in the laboratory on wear and life of various
paper. It was important up to a point but another means of parts. He considered that caution should be exercised before
ensuring freedom from choking was to control the rate of flow taking too much notice of them because they appeared to him
and control the concentration. In that connexion, perhaps, it to relate to a short period, and that it was first necessary to wait
would be necessary to consider the use of servo-mechanisms until tests had been carried out in the field over much longer
connecting the various functions of the pumping plant, the speed periods.
of the pump, the amount of water that was pumped through it, In regard to degradation, he believed he was correct in saying
and the rate of flow of coal into the system, in order to make the that however much within reason the coal was degraded it did
mechanism stable and, at the same time, make it work with the not appreciably lose its calorific value. Therefore, if by pumping
highest possible concentration and the lowest safe velocity. it was possible to reduce the cost of coal to the user, then in
Reference had been made in the paper to the suggestion by the course of time there might be not such a great demand, and
Durand and Condolios (1952) that the engine governor might be not such a high price paid, for the larger lumps.
locked. In his opinion few would agreed that that was feasible, It was, in his opinion, an instance of an old engineering
but nevertheless it might be possible to introduce some form of practice being attacked by means of modern scientific methods.
speed control which could be automatically adjusted to avoid It might well be that, as a result, wide new fields of activity
the possibility of choking and, in fact, in a lot of dredging plant would be opened up for hydraulic engineers and that that would
used at that present time speed control was adopted for that very develop to the ultimate benefit of the whole community. Much
reason. work had already been done, but there was much more to be
The authors had discussed the different kinds of pump which done. Many new devices and systems might be conceived before
could be used and the necessity or otherwise for using a feeder the most economical methods were found. It was essential not
to inject the coal into the high-pressure water. The simplest to be discouraged by the difficulties which presented themselves,
form, and the one which was most commonly used at that present but to use one’s ingenuity to overcome them. It was generally
time, was a centrifugal pump in which the solids passed through understood that, notwithstanding the rapid development of
the pump. Fig. 22 showed a pump which was currently being atomic energy, coal would remain an essential source of power

Fig. 22. Section Through an Ash Pump

(Reproduced from Wauchope 1953, Fig. 16*.)

used for ash. The size of the pipe and the width of the impeller and heating for many years to come. It would be a great achieve-
at its narrowest part would be observed. If larger solids were ment if hydraulic transportation of coal were to give to the
to be passed, it would be necessary to increase that width, and community the benefits of increased availability and reduced
in so doing the efficiency of the pump would be reduced. That costs, and, even more important, if it minimized the grave danger
was one of the reasons why that type of pump would not neces- to health of the presence of coal dust in mines and made the
sarily be very efficient for large solids. The normal maximum winning of coal a more congenial occupation.
head was about 200 feet, so that for high heads it was necessary
to put a number in series, which presented a special design Mr. L. E. PROSSER, B.Sc. (Eng.) ( M e m b e r ) , said that he wished
problem. to emphasize some of the broader economic implications of the
On the question of feeders, it was possible that many more work of the authors.
devices would be invented before the correct solution was
reached, and many more systems might be adopted. What he * WAUCHOPE,
G . A. 1953 Colliery Engineering, p. 415, ‘Moving Coal
had in mind in that connexion was a possible combination of a by Pipeline’.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 DISCUSSION ON HYDRAULIC TRANSPORT O F SOLID MATERIAL I N P I P E S 575
The main value of the paper was that it had shown that it was As mentioned in Bagnold (1954) Sir Geoffrey Taylor had
possible to predict on a scientific basis the engineering require- pointed out to him that, if the effect of those impacts was mainly
ments for particular installations, and hence their running costs the creation of spin, the dynamic-stress ratio should, on
and possibly also their capital costs could be estimated with theoretical grounds, be 0.2 for wholly elastic grains and 0.4 for
reasonable accuracy. wholly inelastic ones, regardless otherwise of the material of the
If a number of installations was examined, it was found that grains.
there were many instances where that kind of transport would When the flow velocity through the pipe was large, the rate
have considerable financial advantage over alternative methods, of shear was large, the impacts became more violent and the
in addition to the other incidental advantages mentioned in the degree of resulting dispersion by scatter became greater. So it
paper. Obviously, there was likely to be an increasing demand for could be assumed that, ultimately, the grains became uniformly
its adoption both in Britain and overseas. It was possible that dispersed over the cross-section of the pipe. At the same time,
in the course of time it would become even more advantageous when the flow velocity was large the backward slip velocity of
because the tendency was for power costs to decrease in relation the grains through the fluid could be neglected. Thus at high
to other costs, and the biggest item in that type of installation flow velocities the discharge ratio, which the authors had called c,
was its power consumption. Therefore, in the future, the could be replaced by the real concentration in the pipe, which
balance was likely to be in favour of that method. the authors had called q.
He feared that unless quick action was taken Britain would Under those conditions, the interesting dimensionless ex-
be forestalled in that field by the other countries. The French pression (i--iW)/c(s--1) became simply the dynamic-stress ratio,
were very active and, no doubt, the Germans and the Americans namely, shear stress attributable to the grains divided by the
were also keeping themselves informed. British engineers might immersed weight of the grains, which he himself had found
experience the mortification of seeing British collieries made by direct measurement of the repulsive pressure normal to the
more efficient by the use of foreign machinery developed from boundary to be 0.3, and which might theoretically be as low
British research, and a large overseas market for similar as 0.2. It would be seen from Fig. 5 that the authors had found
machinery would then also be lost. It was partly for that reason that the ultimate value of (i-iw)/c(s- 1) approached about
that the British Hydromechanics Research Association had 0.25. The agreement was rather striking.
given a lead in the development of the feeders described in the It seemed to him that that ultimate value of the ratio
paper, but if that side was to be developed quickly much greater (i--iW)/c(s-1) was an important experimental quantity. It was
effort had to be made. a measure of the minimum pumping head required. He wondered
It was most desirable that that sort of work should have much why the actual value of that ratio rose so steeply as the flow
greater active support from the British manufacturing industry, velocity was reduced. The rise was obviously connected with the
particularly on the development side. A sum comparable with settlement of the grains into the lower part of the pipe, where,
that spent by the British Government on aircraft development moving more slowly, they accumulated and reduced the effective
should go into industrial development, but there had to be an pipe diameter.
active lead from industry. If that was true, then the conditions at the desirable lower
flow velocities ought to be improved considerably by rifling
Brig. R. A. BAGNOLD (London) said that the finite ultimate the pipe. By imposing a secondary circulation having a velocity
value of the quantity ( i - i w ) / c ( s - l ) , i.e. the excess pressure near the pipe wall somewhat greater than the velocity of fall
head divided by the concentration times the immersed weight, of the grains a much more uniform grain distribution over
was, he considered, of some interest. the cross-section ought to be obtained; and the grains would
The paper dealt largely with the transport of material which virtually be thrown into suspension.
was too large for suspension by fluid turbulence to play any For instance, with the $-inch coal in the 3-inch pipe at a flow
appreciable part. In that event, in a horizontal pipe, it was clear velocity of 3 ft. per sec. (Figs. 4 and 5), the rate of fall of the
that if the grains were not so suspended they had to support grains was given as being about 0.5 ft. per sec. With a rifling of
themselves when travelling as a dispersion. There was no alterna- one turn in thirteen diameters, which was a rifling angle of
tive. Hence, the whole immersed weight of the moving grains 1 in 4 or 0.25, an upward velocity near the wall of about 1-5
in the pipe must be exerted against the pipe wall. In other words, times the fall velocity for the lumps of coal would result. The
the solids must ultimately rest against the pipe wall although grains being in suspension, the ultimate grain-friction head of
they were at any moment dispersed within the fluid. 0.25c(s- 1) would be removed, and for it would be substituted a
The results of his own recent experiments (Bagnold 1954)*, lifting head of c(s- 1) times the angle of the rifling, which was
in which the effects of gravity had been removed, showed that again, he supposed, 0.25. Thus according to Fig. 4 the pumping
when grains dispersed in a fluid were being sheared over one head should be at least halved, even after allowing for a reason-
another, overtaking encounters between grain and grain and able extra loss due to the rifling.
between grain and wall gave rise to what amounted to a repulsive Other advantages seemed to follow, such as reduced degrada-
pressure between them. If gravity then acted transversely tion because suspension would be substituted for sliding and
towards the boundary or wall, that repulsive pressure was impact, and reduced value of the true concentration in the pipe
sufficient to maintain the grains statistically dispersed against due to higher mean grain speeds with, consequently, less liability
the transverse force of gravity, even in the entire absence of to choking and the possibility of still higher discharge rates
fluid turbulence. being used.
That appeared to hold for all cohesionless grains whatever The idea was so simple that there had to be a snag somewhere,
their size, whether or no the effects of their inertia were signifi- but he could not discover where. It had, he believed, been tried
cant relative to those of the inertia of the fluid and whether or once in the United States. He strongly suggested that it should
no the fluid was turbulent. be tried again. He could not see any practical difficulties about
It had also been found that the additional resistance to shear the rifling. He had in mind indenting each separate length of
attributable to the presence of the grains was proportional to the pipe by drawing it through a ring of, for example, six independent
repulsive pressure between them which kept them dispersed. In grooving rollers set on the skew. Those would be brought
other words, even though the grains were dispersed one from inwards to be applied, for example, one pipe diameter behind
another, there was a definite dynamic-stress ratio or friction the head of the pipe and released one diameter before the tail
coefficient, shear-stress/normal-stress, just as there was in the of the pipe came along, so that it would leave one diameter of
static case when the grains were all resting on one another. free round pipe at both ends for jointing.
In that particular instance the grains were large and hard, The implication of that idea was in effect that what should be
their inertia effects were pronounced, and encounters between done artificially for large grains was just what Nature, by
them and between them and the pipe wall were in the nature of turbulence, did for fine particles, namely, supply sufficient lift
impact collisions. Under those conditions it was found experi- to keep the grains in suspension.
mentally that the dynamic friction coefficient was about 0.3.
Mr. E. G . WHITE,Assoc.1.E.E. (Manchester), said that the
* BAGNOLD,
R. A. 1954 Proc. Royal SOC.,vol. A225, p. 49. title of the paper was rather misleading, sincethe subject-matter

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

576 D I S C U S S I O N O N H Y D R A U L I C TRANSPORT O F S O L I D MATERIAL I N P I P E S
did not deal with all types of solid but was related mainly to the
transportation of coal and other large-size solids.
The authors had stated, in the introduction and elsewhere in
the paper, that for hydraulic transport, even for comparatively
short distances, the power requirements were large. He wondered
whether that was true for all cases, particularly where fme-size
solids flowed in suspension as opposed to saltation.
He had been concerned in an investigation at the Laboratoire
Dauphinois d‘Hydraulique, Grenoble, carried out on similar 25 04
uy 0 /
200 u + o o
lines to those described in the paper, in regard to the transporta- CONCENTRATION-GRAMMES PER LITRE
tion of fly ash produced by large pulverized-fuel-fired boiler
plants. It might be of interest if some of the results obtained with Fig. 24. Critical Sedimentation Velocity as a Function of
that class of material were compared with those given in the Concentration for Two Test Pipes of 40.6 and 150 mm.
paper. Diameter
Fig. 23 showed the results of a sizing analysis of the particular Pipe diameter, mm.
-Q- 150
---A,--- 406

Fig. 25 showed the variation of the parameter v , / l / ( g D ) with

concentration, where v, was the critical sedimentation velocity
in metres per sec.
A dispersion took place of the points which referred to the
smaller pipe at high concentrations and it was considered that
that was due to the appearance of plastic phenomena in the
mixture.
Fig. 26 showed the typical operating conditions for a pumping
system deduced from the investigation as under: length of
pipeline, 16,400 feet (5,400 yards) j size of pipeline, 10-inch bore j
total head, 151 feet for steel pipe, 183 feet for plain cast-iron
pipe; quantity of mixture pumped, 972.4 g.p.m. j concentration
of mixture, 3.83 lb. per gal. of mixture; velocity of flow, 4.75 ft.
per sec. j motor power (motor 90 per cent, pump 65 per cent),
76.2 h.p. for steel pipe, 92.5 h.p. for plain cast-iron pipe.
The actual pump characteristic was an assumed curve to
illustrate the desired operating condition of the system.
From the foregoing, it would be appreciated that those con-
siderations were very different from those investigated by the
authors.
Pulverized-fuel fly ash was often highly alkaline and might
contain considerable quantities of highly abrasive solids. He
asked the authors to give an opinion regarding the use of
polythene pipe linings for such a duty.

L
0I 0.5 0.2l 0.1
MIPI.
-
0.05 0.02 U 0.01 ,s U
MIC~ONS
2,
He would be interested to know what effect the presence in
the mixture of sulphur, pyrites, and clay would haveinregard
to corrosion, erosion, and blockages.
SIZE OF GRAIN If coal slurry was allowed to stand in a pipe for a considerable
period, he asked whether the pipe could be easily cleared and
Fig. 23. Granulometric Analysis Before Tests of Fly Ash what would be the effect of gas release from the coal.
from Portobelb Power Station Used in Investigation On p. 568, reference had been made to the pumping system at
Carling. That system included specially designed quick-closing
valves immediately after the pump discharge, to resist reverse-
Curve 1 Mean sample taken from wave effects in the event of a sudden shut-down of the plant.
A Drum No. 21 He asked whether the authors considered that the expense and
B Drums Nos. 20,23, and 26 inclusion of such equipment was absolutely necessary.
C All sixteen drums The authors had referred on p. 571 to a method of concentra-
D Drum No. 19* tion measurement by gamma-ray absorption. He had reason to

* Drum No. 19 contained samples of dust obtained immediately

after soot blowing.

type of fly ash used in the investigation. It would be seen that

the range of size varied from approximately 0.5 mm. down to
about 2 microns, 80 per cent being below 100 microns, and 50
per cent below 20 microns.
The flow of that material, when mixed with water up to a
concentration of 50 per cent, equivalent to approximately 5 Ib.
of solids to 1 gallon of mixture, was entirely in suspension and
6 200 400 660
without any deposition in the pipes. CONCENTRATION-GRAMMES PER LITRE
To determine the pumping velocity, tests had to be carried
out and Fig. 24 illustrated the variation of the critical sedimenta- Fig. 25. Variation of the Parameter v L / 2 / ( g D )as a Function of
tion velocity as a function of concentration for 40.6-mm. and Concentration for Two Test Pipes of 40.6 and 150 mm.
150-mm. pipes. It would be noticed that the critical velocity Diameter
varied little with concentrations above 100 grammes per litre Pipe diameter, mm.
(1 lb. per gal.) but that, for any given concentration, the variation A 406
was appreciable with pipe size. Q 150

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 DISCUSSION O N HYDRAULIC TRANSPORT O F S O L I D MATERIAL I N P I P E S 577
of holding particles in suspension, and these will then contribute
200
very little to the pressure drop along the pipe’. He did not agree
with that statement. The particles were falling continuously by
I50 gravity and were spasmodically being lifted up by eddy diffusion,
and the energy required for that lifting had to be provided by
loss in the pressure of the liquid.
The short section on Flow in Sloping Pipes (p. 566) was,
I00 he considered, most interesting. It was the first attempt to study
90 the flow in other than vertical or horizontal pipes. He wondered
80 f whether it was applicable to all sizes of particle and to the full
70 2
range of concentrations. As he understood it, it meant that when
pumping from one point to another, the excess pressure drop due
60 2 to the solids was independent of the path taken by the pipe. Of
Y
course, the pressure drop due to the water alone would increase
50
with the length of the pipe, but that was usually the considerably
8 smaller of the two components. I n his own experience, the safest
40
way of pumping a slurry to a point at a higher level was first to
b-
ul run the line vertically upwards to a point higher than the final
x one, and then to use a downward-inclined pipe to cover the
30 ) , horizontal distance. That was to reduce the danger of choking.
$ He would be interested to know whether the authors considered
I
that that arrangement might be advantageous sometimes. The
20 authors had demonstrated by means of a film that spherical solids
went round the bend in a pipe very easily, but he wondered
what would happen with rougher particles.
15 The researches on measurement towards the end of the
paper had yielded useful results. It was encouraging to know
that the rate of flow and the concentration of solids could be
10
determined by such simple devices as a Venturi in a vertical
16 I 8 20 30 40 50 60 70 90 100 I20 140 160 180 10 pipe and the measurement of the difference in head over a
FLOW OF MIXTURE-LITRES .R SEC. moderate length of vertical pipe. Since that had been established,
it should be possible to obtain data on hydraulic transport much
Fig. 26. Operating Curves for Installation Using Pipes of more easily and quickly.
10-inch Bore
Mr. A. WRIGHT (Graduate) said that the apparent attractions
of hydraulic transport had been mentioned. So far as the r++g
believe that at Carling that method had been supplanted by industry was concerned, they were its directional flexibility, Its
direct sampling. compactness, and the absence of fire dust and other hazards
associated with mechanical handling plant. In addition, there was
Mr. R. A. SMITH,M.A. (Associate Member), said that the the opportunity for automatic control with a small labour force.
estimation of pressure drop in two-phase flow was a most There appeared, however, to be certain limitations and many
difficuIt problem, but it was essential for assessing the economics important points requiring clarification before large-scale,
of any project for the hydraulic transportation of solids. Measure- economic installations could be attempted with confidence.
ments of hydraulic gradient reported in the paper seemed to be In the latter respect, his organization was fortunate in being
somewhat lower than would be expected from the measurements able to extend the work done by the authors in the laboratory to
made by Durand and Condolios (1952) in the transportation of an experimental pilot plant at a colliery in Scotland. In that
sand and gravel. An exact comparison between the two sets of installation, coal would be raised to the surface from the pit
experiments was not possible because it was not known for bottom, and horizontal, vertical, and inclined pipelines would be
certain how the various non-dimensional groups should be involved. Therefore, a really good idea of the flow in the different
related to each other. In particular, the influence of the density directions would be obtained.
was critical, especially with coal whose density was not much The most important transport problem at a colliery was from
greater than that of water. It would be a great help if research the coal face to the coal-preparation plant. The typical material
workers were to decide on the best method of plotting results. to be handled will be untreated run-of-mine coal with size ranges
If a method were to be agreed, the occasional points obtained down to zero. Particle shape would vary considerably, and there
from industrial installations might then easily be compared. was expected to be a wide variation in the specific gravity of the
Ideally, it should be possible to include pneumatic transporta- solids owing to the presence of shale, stone, pyrites, and other
tion as well. In spite of the much greater ratio of the density of solids. In that respect, it could only be hoped that the formation
the solid to the density of the transporting fluid, the mechanism of ‘dunes’, as mentioned by the authors, could be overcome.
of transportation was fundamentally the same. He had tested In view of the high heads involved with the transport of solids,
an installation for the pneumatic conveyance of cement along the possibility of pipe fracture due to weakening by erosion
a long pipeline. The point calculated from that lay about 25 per warranted serious consideration. Use of special linings to reduce
cent below the curve in Fig. 5 . He wondered whether the authors the frictional resistance would perform the twofold purpose of
would comment on that. The difference might be due to the economy in power requirements and extension of pipe life.
small size of the cement. As mentioned by the authors, there was much speculation in
In hydraulic transport it was known that with large or with regard to the maximum permissible particle size for a given pipe
small particles, the particle size had no effect, or a very small bore. It was obvious that for the economical operation of systems
effect, on pressure drop. However, there was a considerable, and a pipe had to be designed on the basis of the volumetric through-
commercially important, intermediate region, where the size had put of solids rather than the top size, and that would be deter-
a considerable effect. Therefore, it followed that in that region mined by the pipe selected. The effect of such a restriction on
it was essential to understand the effect of size grading, and that the salable price of the coal would require individual considera-
was particularly so when there was a wide range of size, as tion, and underground crushing would invariably be necessary.
mentioned by Mr. White. I n the results reported by the authors, In all except very small installations the head of the coal-
the particles were probably large enough for their size to have transport system appeared to be such that multi-stage turbine
little effect. In the paragraph on Effect of Particle Size (p. 565) pumps and coal-feeding devices would be necessary on the
the rather difficult intermediate region was dismissed rather grounds of system efficiency,power requirements, and, possibly,
lightly with the remark ‘the turbulence in the liquid is capable. degradation of the solids.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

578 DISCUSSION O N HYDRAULIC TRANSPORT O F S O L I D MATERIAL I N P I P E S
Apart from the obvious advantages of a closed circuit for coal dangerous to operate in practice. Prudence suggested that a little
raising, there was the practical advantage of the use of inter- extra horsepower was a cheap form of insurance against the
connecting valves between the two pipelines to flush the system necessity for expending hundreds of man-hours in clearing
clear of blockages. choked pipelines several miles long. He considered that the
Consideration of hydraulic transport for coal raising from a authors had spent considerable time on those low velocities
mine revived the argument for constant delivery at low speed as which were not really of interest to the practical man.
compared with intermittent delivery, higher speeds and high The same applied to the illustrations, except the top one, in
accelerating and decelerating torques associated with conven- Fig. 2. The lower five were certainly not his idea of a safe way
tional winding equipment. However, the former was still in its of conveying solids in a pipeline. If anything went wrong with
infancy and would require tremendous effort and success to the delivery and the material being conveyed was coal and not
bring it to the same state of reliability and efficiency as the plastic, it would, he was sure, be a tremendous job to keep the
modem electric winder. system going.
On the question of feeding devices, all water-displacement He suggested that the next step ought to be to leave the some-
feeders appeared to present serious practical difficulties by dis- what artificial atmosphere of the laboratory and go out into the
charging contaminated water from the system. That feature field to see some of the many installations where a practical job
became more apparent underground where high heads were was being done, and the data should be based on that.
involved at the feeder, and clarification of the water was neces- If the coal-mining industry was to be revolutionized, as was
sary before reintroduction into the pipeline could be carried out. likely, by that method, the National Coal Board ought to be
His opinion was that insertion of solids into the pipeline by prepared to spend a lot of money, not for laboratory work but
means of a specially designed solids-handling mechanical- for full-scale work in the field with real sizes of pipe handling
displacement pump was the solution. Such a pump should be real sizes of coal.
sufficiently robust to handle large coal particles, capable of
dealing with large throughputs and sufficiently sensitive to Mr. J. A. MAGILL(Graduate) said that the Engineering Branch
maintain system stability. of the National Coal Board, in conjunction with the Scottish
A prototype of such a feeder had been designed at the Division, were organizing a full-scale experiment at a Scottish
headquarters of the National Coal Board in London, and it colliery to study the problems of hydraulic transport, and it was
would be tried out at the end of 1955. Operation would be intended to convey minus 2$-inch coal in a 7-inch pipe.
through oil hydraulic jacks, with ample opportunity left for The objects of the experiments were : (1) to find whether coal
servo-control at a later date. The forthcoming trials should could be pumped reliably; (2) to find the precise top size of coal
enable valuable information to be gleaned and should do much that could be pumped in a given pipe; (3) to obtain information
to verify the authors’ predictions. on feeder design, feeders being essential at the heads con-
templated; and (4) to study the integration of that type of
Mr. I?. B. SILK(London) said that ever since, some twenty- transport with the rest of the transport and mining system.
five years previously, he had first introduced the high-pressure The National Coal Board had been accused of hesitation by
sluicing system into Britain, he had been intimately concerned previous speakers and to justify that he would like to indicate
with the pumping of solids in pipelines. the standard that hydraulic transport had to achieve in order to
I n that system, from the ash dump onwards, a common method compete with existing methods.
of handling the material was, of course, by pumping, either to an The Scottish experiment would be in a pit 280 feet deep. Two
overhead bunker or to waste ground up to three, four, or five outputs were to be considered, 100 or 150 tons per hr. A new
miles away. Therefore, it was nothing new to him. In fact, the electric winder was also being considered for the same duty.
sluicing system generally had practically become standard equip- The efficiency of the electric winder would be 45 and 32 per cent,
ment for the large modern power station. respectively. Based on the authors’ methods of calculation, the
Therefore, he had been rather surprised when reading the efficiency for hydraulic transport would be 35 and 40 per cent,
first sentence of the paper to find that the authors had made no respectively.
reference at all to that most important application of the art. As a sample of a more arduous duty, there was a colliery in
In their opening sentence they had referred merely to dredging, Scotland for which a multi-rope friction winder with balanced
land reclamation, or the winning of minerals. For each applica- skips was contemplated. The duty was to raise 286 tons per hr.
tion for such purposes he could show them ten or more applica- from a depth of 2,800 feet. The efficiency of the winder would
tions, and that was not counting the United States of America be 63.5 per cent. On the same basis the efficiency for hydraulic
where there were probably five for every one in Britain. He transport would be 4 8 5 per cent.
estimated that with the equipment that had been made in It was also necessary to consider the mechanical-handling
Britain probably 15,000,000 tons of solids were being handled equipment that must go with the system to load and unload it.
in a year. He did not know what the figure was for the United Fig. 27 helped to illustrate that, and it could be seen that a great
States. That was not in one or two big installations; it was in improvement over a tub-and-cage winding was made by
hundreds of them. Therefore, his company had a very diversified installing skip winding with tipplers and a tub circuit at the pit
experience, and it seemed strange to him that recourse had not bottom only. At its best that system could be entirely automatic
been had to that experience. except for man winding.
The paper was obviously the outcome of much hard work. It appeared that hydraulic transport would require much the
He was most interested to see how closely it followed and con- same handling plant as skip winding, but with the additional
firmed previous papers by Durand (1953, etc.). However, he complication of a crusher and screening plant.
preferred to regard the paper as a preliminary only, because it In that connexion it was relevant to quote the example of a
was on such a small scale-almost a model scale in the laboratory. small pit from which was wound 600 tons a day. That pit had
The great difficulty in assessing scale effects where there were so been converted from cage to skip winding at a cost of about
many uncontrollable variables was well known. E20,000 and with no interruption of coal production.
For instance, what happened in a l+-inch pipe with +-inch It must also be remembered that other factors besides trans-
particles was nothing like what would happen in a 13- to 18-inch port influenced shaft size. For instance, ventilation-air velocities
pipe carrying large coal. Yet in Table 1minimum velocities were must be limited to 1,000 ft. per min. in roadways and 1,500 in
given for pipes of up to 18 inches. The only justification for that shafts. Two shafts of 16 feet diameter were often required for
that he could find was a statement that minimum friction that reason.
increased as the square root of the pipe diameter. He wondered He hoped those points served to illustrate the high standard
whether that had been verified in practice, or whether it was that must be obtained. In the application to coal mining, feeders
based on extrapolation of the curves which had been found from of some kind were essential and for that reason feeder design was
1+-,3-, or 6-inch pipes. It seemed a dangerous thing to do that. a major factor. For the duty he had quoted where 286 tons per
The same was true in regard to Fig. 3 where some curves had hr. were raised 2,800 feet, about a third of the horsepower was
been shown. Much time seemed to have been taken in establish- lost if leakage was 20 per cent of the total volume.
ing the conditions of flow at velocities at which it would be Another major problem was the precise size of coal that could

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 D I S C U S S I O N O N HYDRAULIC TRANSPORT O F S O L I D MATERIAL I N P I P E S 579

a Basic layout. b Cage winding.

6 -

c Skip winding with haulage. d Skip winding with conveyor.

Ld
e Hydraulic transport from pit bottom. f Hydraulic transport from coal face.

Fig. 27. Various Methods of Coal Raising

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

580 DISCUSSION O N HYDRAULIC TRANSPORT OF S O L I D MATERIAL I N P I P E S
be transported in a given pipe. The nearer the system was to there were considerable fluctuations, and that was something
the working face the more adverse did the size/quantity ratio which had to be faced. In large installations, no doubt, the
become, rendering the system less and less efficient. problem could be suitably dealt with.
A third problem was integration with the rest of the mining There was one point on feed sizes which might also be of
and transport system. Many difficulties might be solved if right interest. His company had found that feed sizes should be slightly
at the coal face it were possible to mine the coal in 3-inch cubes. less than half the pipe diameter. There had been a case in point
A fourth problem arose from the cost of cleaning water in a where a screening grid had been used for controlling the feed
closed-circuit system. T o attain 10 per cent of solids in sub-sieve size into a 6-inch pump. It had been operated at 2$-inch intervals.
sizes might not prove too costly but whether that could be It had been considired at first that 3-inch intervals would be
pumped to heads of 300 feet and more was another question. satisfactory but that had not proved to be so and it had been
He wished to ask the authors whether they had eyer used necessary to revert to 23 inches.
fibre-glass tubing and, if so, how it compared with steel tubing
for friction losses. Mr. E. CREWDSON, A.M.I.C.E. (Kendal), said in regard to
the success secured so far by the authors with their reciprocating
Mr. B. G . FLETCHER (London Colney, St. Albans) said that feed that he knew they had not got to the end by any means,
he had a few figures from the point of view of the user or and he was sure they would not claim so, but the success that
operator of pumps with stone and sand, which might be of they had attained so far seemed to him very hopeful, in fact,
interest. It had been said that there was nothing new about that, much more hopeful than the lock hopper with so many flap
but he considered that the speeds and concentrations used by valves and so many more control valves, which seemed to him
his company were, perhaps, a little different from normal prac- rather terrible.
tice. However, he wondered whether the authors were perhaps too
The pumps were all centrifugal, being 8, 6, and 5 inches. preoccupied with aiming at high efficiency. It seemed to him so
Normally, the speed of working was 12 ft. per sec., except much more important to achieve safety of working rather than
occasionally, when it was a little above that. The concentrations high efficiency. An earlier speaker had stressed that point, and
were normally 20 per cent, and those had been checked over a he was strongly in sympathy with him.
number of years, working on yearly averages. That had been At the Carling power station and the Canadian power canal the
done while the difficulties encountered in respect of high wear consumption of power seemed to range somewhere around 12
and tear were recognized, and it was considered economical to kW.-hr. per ton. Assuming that the lift was direct 1,000 feet
do that because certain alterations had been made in the approach from a mine, he had worked it out that about I+ kW.-hr. per ton
to the matter and it had been found that the heavy wear and tear was the 100 per cent efficiency mark. If it was economical at
obtained did not offset the gain that was otherwise achieved. Carling and at the Canadian power canal to use 12 k W . 4 ~ per.
It might be of interest that the pipes which were used for ton, he presumed it would be somewhat similar in a mine where
pumping both sand and gravel with both the 8- and the 6-inch the lift was 1,000 feet. Then the overall efficiency might be
pumps were normally of +%-inch wall, which lasted about something of the order of 8-10 per cent. He considered there
eighteen months. were other devices which could then be used which were simpler.
With the finer materials, such as the sand, which was highly He was thinking of the air-lift method, which seemed so much
abrasive, the pump impellers had, until a fairly recent date, been simpler than any of the other mechanical devices.
lasting a matter of weeks. However, by means of rubber-lined He had one point of criticism concerning the toroidal pipe
pumps a much longer life had been achieved. In one instance, with which the authors had tested the degradation of coal. They
the lining had lasted more than a year. It was, he considered, appeared to consider that the circular pipe would give the same
rather interesting that in that instance a slight difference in the effect as would obtain in a horizontal pipe. He disagreed and
grading of the material had a profound effect on the life of considered that in the section which went up past the horizontal
linings. diameter, the coal and water were travelling vertically in the
I n regard to the control of density, it had been found that pipe.

Communications
Mr. N. BROOK(Leeds) wrote, in reference to the section on and if the actual concentration in the region of the meter were
Measurements, that it was a relief to note that the authors did known, that correction would bring the coefficient of discharge
not claim an accuracy of greater than f 10 per cent. The details nearer to 0-975. If the discrepancy between the coefficient for
of tests with a venturi-type nozzle were interesting and the use the meter when transporting solids and when passing clean water
of mixture density in calculating the discharge was especially so. could be accounted for in that way, the results would confirm the
Experiments were being carried out in the Mining Depart- tests carried out in France, referred to in the paper, in which the
ment of the University of Leeds, using a non-standard Venturi meter (an eccentric nozzle) was horizontal and the differential
of diameter ratio 0.85 for measurement of upward vertical flow pressure was expressed in feet of water.
in a %inch pipe and the limited number of test results so far There was a great need for a simple and reliable method of
available showed that, when pumping crushed bakelite of a size concentration measurement.
range of -&+ inch, the coefficient of discharge of 0.975f0.2 per Experience with the vertical-flow method of concentration
cent was modified to 0.90f0.75 per cent when the meter measurement described ia the paper showed that that method
differential was expressed in feet of water; if the differential was was difficult to apply in small-scale testing. The vertical height
calculated in feet of mixture, the coefficient was then 0.926 *0.69 of pipe used had to be greatly shortened (down to 1 foot for a
per cent. The results of that modification to the readings were 2-inch diameter pipe, compared with the 10-20 feet mentioaed
not 2s encouraging as had been the experience of the authors. in the paper) in order that the frictional pressure loss due to
Another method had been used to attempt to account for the clear water could be compensated by a reasonable amount of
difference in discharge coefficient; in that method a term, contraction, i.e. a contraction which would allow solids to pass
corresponding to the excess pressure difference due to the 2nd not cause blockage. It followed that a sensitive manometer
mixture density, was subtracted from the meter differential was necessary to measure the small pressure difference on rhe
head ;that term was calculated from the vertical height between 1 foot of pipe; owing PO the presence of solids and to the COE-
the pressure tappings and the delivered concentration. That siderable fluctuation in manometer levels which occurred no
modified the coefficient of discharge to 0.929f0.44 per cent; great accuracy could be secured.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 COMMUNICATIONS ON HYDRAULIC TRANSPORT O F S O L I D MATERIAL I N P I P E S 581
He asked whether the authors considered that more attention
should be given to electrical (or electronic) methods of con-
centration measurement or were they hopeful that a reliable
hydraulic method could be developed.
Mr. R. DURAND and Mr. E. CONDOLIOS (Grenoble) wrote
that the authors had made a thorough overall presentation of
the problems arising in the hydraulic transport of coal in pipes.
They would present their own remarks in the same sequence as
the relevant points had appeared in the paper.
Introducing the test material by a supply device downstream
from the pump did indeed allow very stable flow conditions to be
obtained. That resulted from the fact that the system obtained
was what they called an ‘open circuit’ installation, namely, a
circuit in which the material discharged at the end of the test
pipe was not immediately reintroduced into the installation, but
remained for a time in the storage tank. In reality then, it was
not only the fact of introducing the material after the pump which
rendered the system more stable.
The authors had indicated that the accuracy of the measure-
ments was about f10 per cent. They agreed with the authors
that it was difficult to obtain greater accuracy with the large- 2
size material they had used in their tests. But for smaller
material, less than & inch, for example, it was possible to obtain
an accuracy of f3 per cent.
Fig. 28. Head Losses for Coarse Grains
In regard to the pressure gradient along the pipe at very high Curve of 4 = f { v / d ( g D ) } .
speeds, they had not stated, as had been attributed to them by
the authors, that for large solids the difference between the
pressure gradient with material and the pressure gradient with
water alone disappeared if account was taken of the greater
density of the mixture. Smooth, 104-mm. Normal, 105-mm. Rough surface,
Their opinion, based on systematic tests, was as under : pipe; Rhone gravel pipe; Isere gravel, 150-mm. pipe;
type I10 Adour gravel,
(1) For solids smaller than & inch forming a plastic paste type A10
with the water, that difference was exactly compensated by
the density correction because the mixture was homogeneous. 0 2.3 - 5.25
(2) For solids between & and & inch there was practically H 5.25-10.0
no difference, i.e. the grains transported in suspension hardly 0 10.0 -15.5
0 15.5 -20.0
used up any extra energy.
(3) For solids larger than inch the difference became
appreciable and grew progressively with the grain size until
A 20.0 -25.0
1 i
it reached a limiting value, which was about twice that corre-
sponding to the density correction.
Thus, considering a progressive increase in grain size, the Authors’ tests
increase in the mixture pressure gradient, as compared with
water alone, was at first equal to the density correction; the
+ Coal 1.5 inches; sp. gr. 1.4
x Coal 0.5inch; sp. gr. 1.4
increase then tended to disappear, and ultimately grew until it
became twice the density correction.
Correlation of Data on Flow of Mixtures of Large Solids and
Water. First, they had compared their own test points with (1) InJEuence of pipe roughness. During their own tests of
those given by the authors in Fig. 3. 1951 they had tried using a similar method of representation but
To do that they had taken Fig. 6 of the paper they had had had to abandon it.
presented at the Colloquium on Hydraulic Transport of the In Fig. 29a and b the two methods of representation were
National Coal Board in November 1952 (Durand and Condolios given side by side, and they had used their own test points.
1952). On the graph 4 =f ( v / d ( g D ) ) , the test points grouped
In Fig. 28 they had plotted d = (i-iw)/(iwxc) against themselves correctly, while on the graph (i-iu)/C = f(i,) there
(v/z/(gD))x (l/d(s- 1)) where D was the pipe diameter, and s was wide scatter. That was due to the fact that the pipe rough-
the specific gravity of the solids, supposed equal to 1.4 for the ness could vary considerably, and proved that the presence of
coal studied by the authors. The correction term l / d ( s - l ) material had the effect of multiplying the value of the head losses
allowed them to compare tests made with materials of different in plain water by a coefficient which was a function of the
densities. concentration.
It could be seen that the test points for the 13-inch coal i = i&++C)
corresponded well with their own. But those for +-inch coal lay
about a mean curve well above their own. They considered that and not i = i,+KC
that difference might be due to the fact that during the tests on (2) Insuence of the density of the materials. In Fig. 30
$-inch coal the grains had been worn down. Even only slight they had plotted their own test points for materials of different
wear could produce sufficient fine grains to cause a marked densities both with their method of representation and with
decrease in the head losses. They had themselves had great that of the authors.
difficulty in carrying out correct measurements on large materials.
It had been necessary to renew the test material frequently and
to keep a check on the size of the particles transported. Effect of Mixed Sizes of Solids. In the preceding few years
Secondly, they were not in agreement with the authors on the they had studied in detail the behaviour of mixtures containing
method of representation used in Fig. 5 : different sizes of solids. The essential results were presented by

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

582 C O M M U N I C A T I O N S O N H Y D R A U L I C TRANSPORT O F S O L I D MATERIAL I N P I P E S

s-. 3
I

0.5

0
0 0.050 0400 0.150

yxP) 2m

a Curve of (i-iw)/(iwXC) plotted against v / z / ( g D ) . b Curve of (i-iw)/c plotted against im.

Fig. 29. Comparison of Two Methods of Presentation

1 1 I I
1 I 1 :%1
0 10-15 104 (smooth’, 2.25 0.0150
2-30-5.25 104 (rough) ‘ 2.25 0-0187
150 (smooth) 0.0156
x
0 i::: 150 (rough) 0.0183

Diameter of grains, 2.04 mm. ;diameter of pipe, 40.6 mm.

0 Pollopas, sp. gr. 1.5. @ Sand, sp. gr. 2.65. + Corindon, sp. gr. 3.65.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 C O M M U N I C A T I O N S ON HYDRAULIC TRANSPORT O F SOLID MATERIAL I N P I P E S 583
them in a paper to the Congres du Centenaire de 1’Industrie type illustrated by the authors appeared to be that the mating
Minerale en France (Durand and Condolios 1955)*. parts could be machined to tighter tolerances owing to the
The problem was complex but might be considered solved. shorter length of the moving parts over which sealing had to be
If all the grains had a diameter greater than & inch, the mean attained.
diameter of the mixture was that which corresponded to the In feeders of that type it did not seem practicable to use any
drag coefficient for the entirety of the grains. If a portion of the type of soft seal between the high- and low-pressure connexions
grains was less than & inch plastic phenomena came into play. and the quantity of leakage would, therefore, depend entirely
It could be said that the presence of fine grains always had a on the size of the clearances and since
favourable effect on the value of the head losses for the mixture. PB3 W
Q K p L
They considered that the lock-hopper feeder illustrated in
Fig. 12 was without doubt the best device for raising coal to the where Q was the quantity, P the pressure, B the clearance,
surface from mines. W the width, L the length of the leakage path, andp the viscosity
of the fluid, it was obviously of greater importance to keep the
Velocity Measurements. I n their experiencethey had observed clearances down rather than to increase the length.
that for the Venturi throat and the contraction flow-meter the A valve of the type illustrated had been tested for 50 hours’
presence of sand grains made practically no change in the clear- running on minus 1-inch coal. I n that valve the bore of the sleeve
water rating. That applied even when the contraction of the was 2 inches and a 2-inch pipe system had been used. At the
Venturi throat was very small. end of the test run leakage at 250 lb. per sq. in. smounted to
4 per cent of the total flow. It was not practicable to continue
Abrasion and Wear of Plant. It was their opinion that the testing beyond that period in the laboratory.
authors had made an admirable summary of the problem of For larger capacities a development o f the valve might be as
wear and they were in complete agreement with them. shown in Fig. 32 which illustrated one pair of ports only. In that

Mr. J. A. MAGILL(Graduate) also wrote that he had raised

the question of feeder design during the discussion, particularly
in relation to the need to limit leakage at the feeder. In that
connexion it might be of interest to give details of a feeder that
had been designed by the Stoke Orchard Research Establish-
ment of the National Coal Board and which had been tested
1 1 LUBRICATING
YATER INLET 1 INLET

jointly there and by the Engineering Branch. The prototype

of that device was illustrated in Fig. 31. It would be seen that it

8
2, /

HI’GH -
PRESSURE INLET
LU6Rl?ATlNG
WATER INLET

Fig. 32. Development of Valve

-Fig. 31. Sleeve-type Feed Valve
1. Valve body. 6. Sleeve valve. version the sleeve had in fact been cut in half and a straight length
2. Low-pressure input port. 7. Drive shaft. of pipe had been interposed between the two halves, thus in-
3. High-pressure output port. 8. Hard-steel inserts. creasing the effective length of the valve. The object of that was
4. High-pressure input port. 9. ‘Ferebestos’ bearings. to reduce the speed of rotation for a given throughput, thus
5. Low-pressure output port. reducing wear on the moving parts. Three pairs of valve blocks
as shown would be geared together and the inlet and outlet ends
consisted of a valve body (9), inside which a sleeve valve (6) would be connected for driving purposes by suitable shaft and
rotated, the sleeve valve having two ports at each end which gears.
engaged with corresponding ports in the body. The ports of the
sleeve extended over 120 deg. in the plane rotation. The sleeve Mr. P. E. MONTAGNON (London) wrote that he welcomed
valve rotated in water-lubricated bearings (9), and at ports the opportunity to add to the experimental background described
where the valve was likely to have to cut lumps of coal in order in the paper a few comments mostly from the practical aspect
to shut there were hardened steel inserts (8). of operating an experimental installation.
The action of the valve was as followed :When ports (2) and (5) I n the description of the feeder being tested by the Ministry
were opened it would be seen that ports (4) and (3) were closed. of Fuel and Power and the National Coal Board stress had been
When ports (2) and (5) were open a mixture of coal and water laid on the limitation of capacity set by the rate at which small
from a solids-handling pump was fed to the cylinder via port (2), coal would fall into water against the upflow of the displaced
thus expelling any water in the cylinder through port (5). As the water. There was in fact considerably more suction applied to
sleeve valve rotated ports (2) and (5) closed and ports (4) and (3) the loader than the diagram indicated in the direction of assisting
opened. Those connected high-pressure water to the cylinder that flow of coal. So far, that had been entirely gravitational but
through port (4) and the coal and water charge was then expelled further modifications had been in hand for some time to
to the output pipe via port (3). That process was repeated as the provide extra suction from the main pump inlet.
valve rotated and the flow might be made continuous by using The point concerning the need to fill the feeder space as
three synchronized valves each having 120 deg. port openings. quickly as possible with coal was obviously of vital importance.
The main advar,tage of chat type of valve over the reciprocating It most certainly demanded the provision of a higher pressure
diflerence than could usually be provided by gravity so that a
* DURAND, R., and CONDOLIOS, E. 1955 Paper to Coflgrks du separate circulating pump was an advantage. Merely using the
Centenaire de 1’Industrie Minkrale en France, June, DonnCes suction side of the main-line pump limited the pressure difference
techniques sur le refoulement hydrauliqus des matkriaux solides en available and might lead to practical difficulties in providing a
conduite’. suitable strainer if main pressures too high for centrifugal-pump

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

584 COMMUNICATIONS ON HYDRAULIC TRANSPORT O F SOLID MATERIAL I N P I P E S
operation were required. It might well be that an injection the essential simplicity of the process might in certain cir-
(Venturi) type of pump for coal loading would have advantages cumstances be worth a lot in terms of overall economy. The
in that connexion in spite of its inherer,t low efficiency. high power suggested at first sight that short hauls were likely
That sort of difficulty became exaggerated when the overall to be the most rewarding. The curves also had an important
size of plant became important (for example, for operation under- value in helping to determine the best size of pipeline for a given
ground). Small size demanded a very short cycle time and delay duty.
in filling all the available space had to be avoided. The close relation between the friction loss in the pipeline
The curves in Fig. 21 showing the power requirements for when coal and plain water were being pumped suggested that a
various conditions highlighted the large power consumption of fruitful attack on the problem of the reduction of power might be
that method of coal transport, but it should be emphasized that to pay attention to possible alternative pipe linings.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

 585

Authors’ Reply
Mr. R. C. WORSTER and Dr. D. F. DENNYwrote, in reply to of reliable theories on the flow mechanics of mixtures. It was the
the discussion and communications, that the contributions had absence of such a theory that made it impossible to standardize
considerably extended the scope of the paper and shown the one method of presenting test data as suggested by Mr. Smith.
wide interest taken in the pumping of solid materials. The restricted nature of knowledge of the subject was well
Brigadier Bagnold had put forward a number of important illustrated by the case of pneumatic conveying of dry cement
views cn the flow mechanics of solid particles in water. The referred to by Mr. Smith: if the air speed in his pipeline was
constancy of the ratio of shear stress to normal stress proposed about 50 ft. per sec., transverse turbulent velocities would be
by him was verified, at least qualitatively, by the results shown about 3 ft. per sec. Since the free falling velocity of the particles
in Figs. 33 and 34. The first graph showed measured values of was probably of the same order as, and their size so much smaller
than, that of the turbulent eddies, an appreciable part of their
weight would be carried by air turbulence. Fig. 5 referred only
to the type of mixture flow where the whole weight of the
particles bore on the bottom of the pipe. That aspect had been
dealt with in some detail by Bagnold (1955)*.
The suggestion that the energy dissipated in turbulence by
the falling component of the particle motion was simply additive
to the energy dissipated in fluid motion alone was not acceptable
to them, and experiments in the related field of open-channel
transport of solid particles in a turbulent bed showed that
the presence of the particles could greatly reduce the fluid
turbulence components (Vanoni 1946)t.
Mr. Durand and Mr. Condolios had pointed out that the type
of correlation given in Fig. 5 differed from theirs (Durand and
Condolios 1952), and had then proceeded to compare the test
data on the flow of coal mixtures given in Fig. 3 with their
Fig. 33. Variation of the Concentration Actually Present correlation (Fig. 28). The resulting discrepancy was one of the
in the Pipe very reasons why they (the authors) preferred to show :heir
+-inch coal; 3-inch brass pipe. results as in Fig. 5. The two correlations were in fact identical
Delivered concentration, so long as the hydraulic friction coefficientf remained constant.
per cent Mr. Durand (Durand and Condolios 1952) had shown that the
+ 13-17 mean curve in his correlation was represented by
e 18-22
0 23-21

and since i, = f V2/2gD it could also be written as

Thus, results on pipelines having the same friction coefficients

f which gave a single curve in Fig. 29a should equally well give
a single curve in Fig. 29b. Yet Mr. Durand’s sets of results on
104-mm. and 150-mm. pipes having equal friction coefficients
iu/ts-i) showed a great dispersion in Fig. 29b. That scatter in the test
Fig. 34. Approximately Constant Relation Between (i-iw) results was equally present in Fig. 29a but it was masked by
and the Concentration Actually Present in the Pipe using a graph whose ordinates varied through a range of 1-8,
while in Fig. 29b the ordinate varied over a range of about 1-2.
Comparison should be made with Fig. 5. $-inch coal; 3-inch brass pipe. The same phenomenon was apparent in Fig. 30 where the test
Delivered concentration, results all came from the same pipeline so that a correlation in
per cent Fig. 30a was identical with the one in Fig. 306; again, the
+ 13-17 dispersion of the results was much more apparent in Fig. 30b
e 18-22 and it clearly showed that the least pressure losses were caused
0 23-27
by the hardest particles (carborundum) and the greatest pressure
losses occurred with the softest (plastic) particles. It was their
own opinion that the unknown influences such as those were of
the ratio of the concentration present in the pipe q to the con- greater importance than the small variations of hydraulic
centration in the delivered mixture c plotted against the same friction coefficient to be found in practical pipelines for trans-
abscissa as in Fig. 5. Those results confirmed his suggestion that porting large solid particles. T o show results so that scatter was
the concentration in the pipe increased greatly at reduced not compressed was the best way to discover the other factors
velocities and that was an impoftant factor in the choking of that influenced pressure loss.
pipelines. Fig. 34 was similar to Fig. 5 except that the excess The views of Mr. Durand and Mr. Condolios on the influence
pressure loss was related to the concentration in the pipe and of mixed sizes of particles on the pressure losses were of great
not to that in the delivered mixture. As Brigadier Bagnold had value because, as Mr. Smith had pointed out, when relatively
pointed out, that ratio was related to the ratio of shear stress to
normal stress in the moving bed and, as he had suggested, it was * BAGNOLD, R. A. 1955 Proc. I.C.E., vol. 4, No. 1, p. 174, ‘Some
not influenced greztly by the speed of flow. Flume Experiments on Large Grains but Little Denser than the
Measurements of concentratiocs, local velocities, and shear Transporting Fluid, and their Implications’.
stresses inside a pipe carrying mixtures were very difficult to t VANONI,V. A. 1946 Trans. Am. SOC.Civil Eng., vol. 111, p. 67,
carry out but they were essential to the development and proof ‘Transportation of Suspended Sediment’.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014

586 AUTHORS’ REPLY O N HYDRAULIC TRANSPORT O F S O L I D MATERIAL I N P I P E S
fine material had to be pumped it would seldom be all of the laminar but with a velocity given not by the Poiseuille formula
same size. but by that of Buckingham (1921)* :
Perhaps Mr. Silk had slightly misinterpreted the paper which
was intended, not so much to describe what was common
practice, but to help in extending that practice. They had not
claimed that ‘the minimum friction loss increases as the square
root of pipe diameter’. In fact the minimum friction loss per foot
of pipe tended to be constant for all pipe sizes when delivering The flow would become turbulent at a velocity which was rather
the same material at the same concentration. It was the velocity uncertain but appeared to be given approximately by the position
at which that minimum occurred that increased as the r o a of on a Reynolds number graph where the value of the friction
pipe diameter. coefficient for laminar flow had fallen to that for turbulent flow.
They agreed with Mr. White that the flow of suspensions of Hydraulic calculations had shown that the pumping of coal
very fine materials such as fly zsh was very different from that of along pipelines would often be attractive, but the machinery
large material such as 1-inch coal. They had themselves also available at that time was generally not suitable for high pres-
carried out tests on the flow of fly-ash slurries which had be- sures. The tests reported by Mr. Magill showed that the rotary
haved in a way similar to that reported by Mr. White : at high plug valve was suitable for application in that field. Mr.
velocities the mixture flowed like a simple liquid but below a Montagnon had described how the use of suction to accelerate
critical velocity the flow was laminar and the pressure loss the coal into lock-hopper types of feeder could greatly increase
tended to a constant value. Their opinion was that that critical their throughput, so that advantage might be taken of the
velocity with very fine materials was determined not so much inherently low leakage of that type of machine. Mr. Magill had
by settling of the suspension as by its viscous properties. If a shown the importance of power efficiency in colliery applications
suspension was tested in a viscometer and found to be a Bingham of hydraulic transport and it was gratifying to see that the
plastic with a yield strength T~ and a plastic viscosity ppl, it development of suitable machinery was being so actively pursued.
would not flow in a pipe until the pressure gradient exceeded * BUCKINGHAM, E. 1921 Proc. Am. SOC.Testing Materials, p. 1154,
Po/L = 47010.Above that minimum pressure the flow would be ‘On the Plastic Flow through Capillary Tubes’.

Downloaded from pme.sagepub.com at University of Leeds on August 19, 2014