E X HIB I T A

How the Sys I ezrz W o z-ks

Acknowledgements Management.
The financial support and confidence of Utah Copper
The successful implementation of SMAP would not have been Division and Kennecott's Metal Mining Division management
possible without the support and efforts of the Utah Copper also contributed to the successful implementation of SMAP.
Division Smelter Maintenance Department and Plant

A New Launder Design Procedure

H. R. Green, D. M. Lamb, and A. D. Taylor

Introduction consistent results.

Second: Provide a rational basis for examining and
The design of slurry launders has usually been based on utilizing operating data to refine and improve the system.
strictly empirical concepts. An examination of the most To accomplish this, it was necessary to develop a basic design
common procedures reveals that they do not account for many concept. This concept has been developed through a complex
of the variables that are recognized as significant for slurry development history, and yet still appears workable and
transport. These may include flow rate, volume concentration technica8y sound. That concept may be outlined as follows.
of solids, solids specific gravity, solids size distribution, particle First: The solids transport velocity is the fundamental basis
shape, launder geometry, and roughness of the wetted surface. for slurry )aunder design.
It was decided to develop a design procedure, which would Second: The stream configuration, hence the launder size,
accomplish two things: is integrated with the solids transport velocity so that the actual
First: Take into consideration most of the known stream velocity exceeds the solids transport velocity.
significant variables and systematize the procedure to assure Third: The launder slope is that which will achieve the

^ 210 AUGUST 1978 SOCIETY OF

required actual stream velocity. investigators. The mathematical treatment employed in this
It was originally planned to utilize a series of charts and correlation reduced the scatter considerably and minimized the
nomographs for the design procedure. A nomograph, based on effect of particle size and shape. The apparent viscosity-volume
data given in Taggart's Handbook of Mineral Dressing, was concentration curve resulting from Thomas's work is used by
developed to d9termine the solids transport velocity. The many designers of slurry pipe lines where the particle size
velocity thus determined was used with Manning charts to distribution is controlled.
determine the launder slope and configuration as though the An apparent viscosity-volume concentration curve for
fluid were water. suspensions of 20 micrometer particles was produced from
It was planned to utilize a ’slope adjustment factor’ to unpublished data. The shape of this curve was similar to
increase the slope to compensate for the apparent viscosity of the Thomas's but exhibited a steeper slope as the concentration
slurry. All attempts to develop a suitable correction factor increased. Further data from a USBM paper by Schack, Dean,
failed. This failure was principally due to the complex and Molloy produced curves for other particle sizes. These,
relationship of viscosity to the required slope. It was then also, had slopes steeper than the Thomas curve in the high
decided to revise the entire approach. concentration region.
Specifically, the present procedure is based on the Darcy- On the basis of these data and from examination of several
Weisbach flow equation, which is rationally preferable to the other curves for apparent viscosity of solid -liquid suspensions,
Manning equation. The Camp equation is used for the solids the Thomas curve was adjusted to accelerate the slope increase
transport velocity. The necessity of reference to the Moody as the solids concentration becomes greater. This adjusted curve
curves for the Darcy friction factor is avoided by using the was assumed to apply to suspensions having particles averaging
Colebrook and White equation. The resulting mathematics 320 micrometers (50@ wt. passing). A family of curves between
require iterative solutions for several of the unknowns. the adjusted Thomas curve for 320 micrometer particles and the
curve for 20 micrometer particles was then constructed. The
Mathematics intermediate curves were developed as a function of the ratio of
increased particle surface area as the particles became smaller.
The discussion which follows concerns the mathematical basis These curves are shown in the Fig. 1 and are used to determine
of the launder design procedure for pipe, rectangular, and U- the apparent viscosity of the slurry.
shaped launders. The equations used are shown in Table 1; the Effective roughness values, Table 3, for various launder
symbols and units in Table 2. surfaces was derived from a family of curves published by the U.
Many papers have been published regarding the apparent S. Army Engineers in 1964. 6
viscosity of solid-liquid suspensions. The experimental data The mathematical relationship given in the above paper is as
which formed the basis of these papers exhibit considerable follows:
scatter even though the experiments were carefully conducted.
D. G. Th 4 correlated data from a large number of
0. 1667
" 23.85 + 2t .95 log (JI/£)

ti — Manning roughness coefficient; It = Hydraulic radius,

feet; I = Effective roughness, feet.
The above equation was solved for ’ b’ and a tabulation of
”1’ values for various ”ri" and ’n ’ values was produced on the
HP9820A calculator.
For any particular hydraulic radius a correlation exists
between the Manning ’ti’ and the roughness value “1’. It is
assumed an average hydraulic radius of 0. 183 m would apply to
launder sizes most likely to be encountered. Errors introduced
by this assumption are minimal. Published ’n’ values for
various types of surfaces and the derived n k correlation resulted
in the approximate "fi’ values shown in Table 3.
The solids transport velocity Eq. I was derived from an
equation by Camp.* Camp's equation was for the self-cleaning
or sediment transport velocity in sewers. The equation as

H. R. Green, Member SME, is Engineering Supervisor, D. M.

Lamb, Member SME, is Engineering Specialist, A. D. Taylor,
Member SME, is Chief Metallurgical Engineer, Bechtel
Corporation, San Francisco, CA. Preprint 7BB16, AIME Annual
Meeting, Denver, CO., February 1978. Manuscript, November 3,
1977. Discussion of this paper must be submitted, in duplicate,
prior to November 30, 1978. In accordance with the Postal
Service Regulations, this material has been assessed a page
charge and is considered advertisement for postal purposes.

MINING ENGIN EE RING 1 211

originally expressed by Camp was as follows: The Froude number, Eq. 14, is the ratio of the flow velocity to
the velocity of elementary free-surface waves. When the Froude
number is greater than one, the flow is said to be supcrcritical,
when equal to one it is critical, and when less than one it is
subcritical flow. Launders should never be designed with
Froude numbers near one. When flow velocities are close to the
P = Velocity, ft/sec; g = Acceleration of gravity, ft/sec}2
wave velocity, dist urbances can cause extreme wave action and
Specific gravity of particles; Dg — Particle diameter, feet; / =
large variations of flow depth.
Darcy friction factor, dimensionless; B — Particle motion
The Darcy-Weisbach friction factor is usually determined by
constant, dimension less.
reference to a family of curves based on the work of F. L.
Camp stated that, when constant "B” is 0.04, the calculated Moody. For more efficient use of programmed calculators and
velocity is that which will start motion of the particles. When computers, the Colebrook and White equation, Eq. 1â, is used
constant ”B" is 0. 8, the calculated velocity is that which will to calculate the Darcy friction factor.
achieve adequate self cleaning in sewers. The solids transport Slope and velocity, Eq. 16 and 18, respectively, are derived
velocity in this paper uses 0.8 as the particle motion constant. directly from the Darcy-Weisbach flow formula. Four times the
Continuity velocity, Eq. 2 through 5, is derived from the hydraulic radius is substituted for diameter to adapt the
fundamental continuity equation Q_ = VA and the geometry of equation to open-channel flow. The Darcy-Weisbach flow
the stream cross section. equation is usually stated:
Hydraulic radius, Eq. 6 through 9, is developed from the
stream geometry and the definition of hydraulic radius. in consistent units
The mean hydraulic depth, Eq. 10 through 13, is used to D 2g
calculate the Froude number. The equations listed are derived
from the stream geometry and the definition of mean hydraulic h — Head loss; f — Darcy friction factor, dimensionless; ñ =
depth. The mean hydraulic depth is equal to the area of the Pipe length; D —— Pipe diameter; P = Flow velocity: g —
stream cross-section divided by the width at the water surface. Acceleration of gravity.

TABLE 1
EQUATIONS

Equation
Number Units Launder Shape Equation
( 6.4g D (S,-11 I’ '
Solids Transport Velocity m/s Pipe, Rectangular
and U-shaped f°
4Q
2 Continuity Velocity mls Pipe
D'I cos '(1-2n)—(2—4nXn-n’P’]

3 Continuity Velocity m/s Rectangular

4Q
4 Continuity Velocity m/s U-Shaped, nd 0.5 V =
W'[cos '(1-2nJ-(2-4n)tn-n'P’]

5 Continuity Velocity mls U-Shaped, n> 0.5 V =

Who/8 + n - 0.5)
D [ cm*'(1-2n1—(2—4n1(n- n')° ’]
6 Hydraulic Radius Pipe R,=
4 cos ' (1-2n)

7 Hydraulic Radius Rectangular nW

(2n + 1)
8 Hydraulic Radius U-Shaped. nd 0.5 R‘ - W[cos*'(1 -2n)—(2—4n)(n-n’P ’]
4 cos*‘(1—2n)
g _ W(a/8 + n - 0.5)
9 Hydraulic Radius U-Shaped, n 0.5
' (r/2 + 2n - 1)
'(1-2n)-12—4nln— n’P ’]
10 Mean Hydraulic Depth Pipe
8(n-n°P ^

1J Mean Hydraulic Depth Rectangular nW

WIcos—’ii-2n)—12-4n) in-n'P’1
12 Mean Hydraulic Depth U-Shaped, nm 0.5

13 Mean Hydraulic Depth U-Shaped, n > 0.5 dp = Win/8 + n - 0.5)

14 Froude Number Pipe, Rectangular,

and U-Shaped
k + 2.51
= -2 log
15 Darcy Friction Factor f’ 14.8R Ref'
S= I (VJ'
t6 Launder Slope mlm Pipe, Rectangular, 4R 2g
and U-Shaped
1IXi0 Q S$
17 Reynolds Number Pipe, Rectangular, R,=
and U-Shaped
(4R, 2g SQ
18 Velocity as a function mls Pipe, Rectangular,
of R#, 5, A f and U-Shaped

SOC IETY O F
1 21 2 AUG UST 1978
Reynolds number, Eq. 17, is stated in one of its many familiar Procedure
forms. This one, again, substitutes 4 times the hydraulic radius
for diameter. The initial launder size is calculated so that the continuity
(design) velocity is 1. 15 times the solids transport velocity. This
Discussion 16 % safety margin is estimated to c ver input data inaccuracies
and launder construction variation.
Having selected the mathematical basis for the flow This design system permits the evaluation of the flow data
calculations, it is appropriate to consider the limitations of such associated with the individual variation of physical dimensions
a basis. The formulae are stated in forms which apply to open- and flow rate. An initial set of flow and launder size data is
channel flow of liquids. It is assumed that the slurries behave calculated from usually available information concerning the
essentially as true liquids. The calculation system does not apply launder performance requirements. This initial set of data is
to solid-liquid suspensions with solids concentration and particle used as a reference or starting point for the eventual design.
size such that the slurry exhibits plastic or thixotropic Based on the initial set of data, the calculated launder size is
characteristics. For pipe launders the flow-depth-to-diameter then adjusted to dimensions consistent with construction
ratio must be less than one. tolerance and practice. Entry of this new dimension creates a
The variety of conditions encountered in open-channel flow is new set of flow data for analysis.
greater than in pipe flow both because of the existence of the After a satisfactory set of flow data is obtained by entering
free surface and the alternative stages of flow having equal new dimensions, the slope is adjusted to conform to construction
energy. Because of this, it is reasonable to expect that the tolerance. Entry of a new slope generates a new set of flow data.
hydraulic transport of solids in launders would be subject to Following a satisfaCtory slope selection, the effect of varying
more uncertainties than in pipes flowing full under pressure. flow rates in the now-fixed launder can be evaluated. This is
This launder design system specifies that the Reynolds done by entering the desired flow rates in the calculations. For
number must be greater than 5000 and that the Froude number each flow rate entered a new set of flow data is generated.
must not be between 0.8 and 1.6. These re.quirements will The adjustment of diameter, flow depth to diameter ratio,
assure adequate turbulence and avoid the unstable how
conditions associated with near critical flow.
In launder design, as in many fluid flow problems, there is no
absolute single answer. Many combinations of Slope, velocity, TABLE 3
and physical dimensions may satisfy the requirements of a
particular problem. The nature of open-channel flow, with its AVERAGE EFFECTIVE ROUGHNESS
free surface, extends the conditions that will satisfy a particular
problem over that of pipes flowing full under pressure.
The mathematical basis and the design procedure apply only Launder Material k, Meters
to uniform flow sufficiently remote from junction boxes, feed
points, and other transitions to produce equilibrium conditions. Concrete pipe:
Lined . gg7$g
Unlined .001350
Asbestos-cement pipe 001350
Clay drain tile, plain .001350
Vitrified clv y tile .001350
TABLE 2
Wood stave pipe .0£D728
NOMENCLATURE
Steel pipe:
Plain .000728
Rubber lined .001350
Symbol Units Description Plastic lined .000350
cos‘' rad Arc Cosine Plastic pipe:
D m Launder diameter Welded joints .000146
D Particle diameter Flanged or coupled joints .002292
Mean hydraulic depth of flow
Froude number Fibreglass pipe (fRP), flanged or coupled .002292
Darcy-lVeisbach friction factor
8 mls' Acceleration of gravity (9.8 mJs°) Cast iron pipe, uncoated .000728
Effective roughness of launder surface
Flow depth to launder diameter ratio Concrete construction:
m'/s Flow rate Formed only, no finish .003627
R, Reynolds’ number Float finish .002292
”h Hydraulic radius Trowel finish .001350
m Jm Launder slope Pre-cast, rough joints
Specific gravity of pulp .003627
Specific gravity of solids Wood construction:
Pa.s Apparent viscosity of pulp Smooth planed surface .0(J£i728
Flow velocity as a function of R„ s, & f Rough surface .005416
m/s Continuity velocity
mls Solids transport velocity Steel construction:
Launder width Welded sections, unlined .001350
meter Rolled sections, unlined
second .000728
Pa Rubt›er lined .0013X
Nm’ pascal Plastic lined .000728
rad radian
N kg.m/s’ newton Pla8tic construction, free formed .000728

MINING ENGINEERS
 ?R0)ECT XYZ EKPAf6t0l
EhCIUEER: hRC,DHL L*UflDER NO: 004

10300
*FLOV OEPTH T0 L*UNDEX 0I*brTER R,TI0

103
2 . gpp

SLURR¥ SOLtOS C0FOEDTD*TI0X

33.00

t. 9

PLOV L*UND¥R D xecy

D*T* ID6flTIPlC*TI0fl

. 0613 6 . 65 1 9. 2 5 . 6 3 I * 06 . 022 I I . 0I +i ï C+i

.08?0 7 . 61 d.69 .600
. 1 250 8 . 7ó
8.32 &.58 . ‹6T
7 . 73
•1Z50 5.37 . 319

.t250 6•97 J.3).352

. i 250
• 1 25O 8. 32 &.58 .320 PLOW R*IE *DJ.
. y y50
. 5 63 SLOPG *D3 STHINT
. G 5 6 3. 9. 0 6
.1563

and slope may be used to bring the Reynolds and Froude and the flow-depth-to-launder diameter ratio is calculated. The
numbers to the desired values. Usually any adjustment that will sequence of calculations for this adjustment is as follows:
result in a higher flow velocity will also result in larger Reynolds 1. Assign and input desired value for the launder slope.
and Froude numbers. 2. Assume a value for the flow depth-to-diameter ratio.
The above sequence of design steps, except the reference one, fl. Calculate continuity velocity (Eq. 2, 3 , 4, or 5,
can be done in any order desired but the sequence noted is the depending on launder geometry).
usual one. 4. Calculate hydraulic radius (Eq. 6, 7, 8, or 9, depending
The calculation sequence for the initial launder size and flow on launder geometry).
data is: 5. Assume a value for the Darcy friction factor.
1. Assume launder diameter. 6. Calculate velocity as a function of hydraulic radius,
2. Calculate continuity velocity (Eq. 2, 8, 4, or â, slope, and Darcy friction factor (Eq. 18).
depending on launder geometry). 7. If the continuity velocity equals the velocity calculated in
3. Calculate hydraulic radius (Eq. 6, 7, 8, or 9, depending step 6, proceed to step 8; if not, assign a new value to the flow-
on launder geometry). depth-to-launder diameter ratio and repeat steps 8 through 7.
4. Calculate Reynolds number (Eq. 17). 8. Calculate the Reynolds number (Eq. 17).
5. Calculate Darcy friction factor (Eq. 15) by trial and 9. Calculate Darcy friction factor. Solve Eq. 15 for ”f’,
t•y y. friction factor, by trial and error.
6. Calculate solids transport velocity (Eq. 1). 10. lf the value of the Darcy friction factor assumed in step
7. If the continuity velocity is not equal to 1.15 times the 6 equals the friction factor calculated in step 9, go to step 11 ; if
solids transport velocity, assume another value for the launder not, use the friction factor value calculated in step 9 as the
dimension and repeat steps 2 through 7. If the continuity assumed value in step S and repeat steps 6 through 10.
velocity equals 1. 16 times the solids transport velocity, go to step 11. Calculate the mean hydraulic depth (Eq. 10, 11, 12, or
g 18, depending on launder geometry).
8. Calculate launder slope (Eq. 16). 12. Calculate the Froude number (Eq. 14).
9. Calculate mean hydraulic depth (Eq. 10, 11, 12, or 13 , 13. Calculate the solids transport velocity (Eq. 1).
depending on launder geometry). 14. All values based on the slope adjustment have now been
10. Calculate Froude number (Eq. 14). determined.
11. At this point all values for the initial launder size and At this point the design of the launder is complete. If it is
flow data have been determined. required to determine the operating conditions at other flow
The calculation sequence for launder size adjustment is rates, the calculation sequence noted for slope adjustment is
nearly identical to the initial sequence. The new diameter and used. The diameter and slope are held constant and the desired
flow-depth-to-launder diameter ratio is used in calculations 2 flow rate is assigned as step 1. Then repeat steps 2 through 18 as
through 6 and 8 through 10 in the same manner as for the initial above.
calculations. Step 7 is omitted. Design summaries of a typical problem for pipe, U-shaped,
The slope adjustment calculation sequence is different from and rectangular launders are reproduced as Tables 4, 5, and 6.
the first two described above. In this sequence of calculations These summaries were produced using Bechtel's Fortran
the launder size and flow rate are fixed. The slope is specified program on a time-share computer.

SO C IETY O F
•=a aMcllST ]P78
Conclusion the-art of slurry launder design. Further refinements, that are
anticipated as the system is used, can result in a better and
While the procedures described sound slightly formidable on better tool for the engineer charged with slurry launder design.
first reading, familiarity with them will demonstrate that the
principles are straightforward and logical. References
The calculations are extremely tedious to complete with a
hand calculator. Bechtel has written programs for the Hewlett l Taggart, A.’ F. , Velocities in Tailing Launders' , Handbook
Packard HP 9820A programmable calculator and also in o Mineral Dressing, John Wiley, New York, 1945, p. 20-55.
Fortran for use on a time share computer. Use of these programs *Camp, T.’ R. , Minimum Velocities for Sewers', Rommel of f/te
has demonstrated that the necessary design calculations are Boston Soci!et y af Ctñi'l Engiheers, Vol. 29, 1942, p. 286.
accomplished quickly and easily. King, H. W. , Wisler, C. O. and Woodbum, J. G. , H ydiautics,
The Bechtel programs are confidential, but the mathematics 5th ed. , John Wiley, New York, 1948, p. 191 .
presented in this paper will allow anyone with similar facilities to 4Tbomas, D. G. , "Transport Characteristics of Suspensions:
produce his own programs with little difficulty. VIII A note on the Viscosity of Newtonian Suspensions of
This procedure was first programmed for the Hewlett Uniform Spherical Particles" , Journal o/ Colloi'd Science, Vol.
Packard HP 9820A programmable calculator. Nearly all of the 20, 1965, pp. 267-277.
available memory (429 registers) was used for a program 5Schack, C. H. , Dean, K. C. , and Molloy, S. M. ,
covering one launder geometry. Those who plan to program this ”Measurement and Nature of the Apparent Viscosity of Water
procedure will find that a machine with more memory will be Suspensions of Some Common Minerals", Rl5334, U. S. Bureau
advantageous. There is, of course, no problem with memory of Mines, May, 1957.
shortage if a shared time computer is utilized. 6 ’Hydraulic Design Criteria ", U. S. Army Engineer Waterways
There has not been sufficient operating experience with Experiment Station, Vicksburg, Miss. , 1964.
launders designed according to this procedure to guarantee the Perry, R. N. , Chilton, C. H. , and Kilpatrick, S. D. , Chemical
accuracy of the individual factors with a high degree of Engineers Hand book, 4th ed. , McGraw-Hill, New York, 1963,
confidence. The system, however, is amenable to optimization, p. 8-299.
without major revision, as further data become available. * Locher, F. A. , ’A Guide to Hydraulic Design of Launders“,
The described design procedure is an advance in the state-of- Bechtel Corporation Internal Paper, June, 1975.

1 +d n AUGUST 1978
SOGIETY OF