Slurry Rheology Influence on the

Performance of Mineral/Coal Grinding

Circuits– Part 2
Richard R. Klimpel

Part 2 of this articie continues the dis- grinding tests, one might logically even at long grind times, thus
cussion of a 10-year, muitimiiiion doiiar expect that dramatic changes in hindering the normal theological
research and piant testing program on any one of the four factors will be transformation presented earlier.
slurry rheoiogy and grinding circuits. The less likely to occur. It will be A final condition that can cause
first part of the articie (ME, Dec. 1982) pre-
shown that continuous mill opera- the occurrence of region B to be
sented the basic concepts identified by
the research and some laboratory test
tions offer some unique opportu- small or zero is when the media
resuits. This section illustrates typicai in- nities to take advantage of possi- void volume filling of slurry is
dustrial scaie test reauits and identifies ble theological transformation by < 10070.
some industrial operating implications of more direct operational control of Figures 5a and 5b use previ-
controlling rheoiogy by different meth- the settings for the four factors. ously published data of this study
ods. One extra observation noted in to demonstrate the influence of
At least four controllable fac- the theological studies was the solids loading and weight percent
tors decide the theological char- variability in the location and ex- solids on the net production oftac-
acter of a slurry-slurry density or tent of region B (pseudoplastic be- onite ore in a laboratory batch
percent solids, particle size distri- havior) for the various coals and ball mill. In particular, Fig. 5a
bution, chemical environment, ores tested. The location of region illustrates several trends not gen-
and slurry temperature. The sec- B was usually in the region of erally recognized until this study.
ond factor has two interrelated 45-55% solids by volume and was They include these two:
facets: the shape of the particle of the extent of O-8%,or 2-ll% with . Increased slurry density al-
size distribution which controls chemical addition. The corre- lows for increased solids loading
packing behavior of the solids, sponding increase in net produc- before passing through the maxi-
and the fineness of the distribu- tion ranged from O-107Oin region B mum in the net production curve
tion. Finer particles increase and 0-2170 in region B’. When re- where the fall-off is due to non
interparticle forces and viscosity. gion B is small or zero (no pseudo- first order breakage.
As indicated in Part 1, during a plastic character is exhibited) no . The use of rheology control
given grinding test it 1s possible increase in production will be ob- ch.emicais such as GA4272 allows
for all four factors to change. How- served and the use of chemicals is this trend to be extended to higher
ever, regardless of the particular often marginal. There are several slurry loadings with an increase
settings of the four factors in a reasons for some materials exhib- in net production over any previ-
given test, if the resulting theo- iting this quick transformation ous condition by keeping grinding
logical character is either dila- from dilatancy to high yield values first order.
tant, pseudoplastic, or pseudo- often at surprisingly low percent Figure 5b shows the same data
plastic with yield, the associated solids by volume such as 3W0. One as Fig. 5a plotted for constant
breakage rate is correlated with condition identified was for ma- weight loadings. These various
the current theological character. terials containing high levels of figures of net production versus
It is obvious, for example, in viscosity—producing elements percent solids show the location
batch grinding tests run at con- such ascarbonates or clays. A sec- and extent of the regions A, B, and
stant percent solids, that the sec- ond condition documented was C presented in Part 1 as a function
ond and third factors, where ap- for materials that exhibited unu- of solids loading.
propriate, are changing during sually fine primary fragment dis- It is obvious that the theological
grinding because size distribu- tributions (fine Bi,j curves). This transformation pattern described
tions are changing and the pro- corresponds to materials having earlier does not hold true for
duction of fresh surface area takes small y values of < 0.5. Such a slurry loadings corresponding to
up unabsorbed chemical. Thus slurry developed a yield value less than the void volume of the
the corresponding theological quickly during grinding because media, which is also a region of
character change in batch tests of the rapid buildup of fines. A re-
with increasing grind time would lated problem can be presented
be dilatant to pseudoplastic to by materials exhibiting exces- Richard R. Klimpel is a research scientist
pseudoplastic with yield. The de- sively coarse primary fragment with Dow Chemicai Co., M.E. Pruitt Re-
gree to which this transformation distributions, Y values > 0.9. search Center, Midiand, Mi 48640, acd a
occurs depends on the changing Breakage of this type of material professor of mineral processing (ad-
setting of the four factors over produces size distributions that junct) with The Pennsylvania State
grinding time. In continuous give poor packing efficiencies University.

MINING ENGINEERING JANUARY 1983 21

1
considerably less than overall
maximum net production. This
underfilled condition is accom-
panied in part by breaking actions 600 wt. % GA-4272
that have a tendency to form ex- ● 72 no
cessive amounts of fines that lead o 77 no
quickly to the formation of yield o 80 no
slurries. It is also worthy to note b 83 no
that the overall maximum net pro- a 83 0.27kg/t
duction without aid occurs in a
weight loading region near that of 100% void
Iwo media void volume loading. volumeloadins
For higher loadings, the transfor- In indicated
by
mation of regions A to B to C oc- 3 500 r
In
3
curs at essentially the same weight
.-c
percent solids values. Only the E
height of the net production ver-
sus weight percent solids curve
changes.
Torque measurements also have
been reported for the data of Fig.
5. It is clear that, in batch tests, the Wat max net production
maximum net production of fines for various wt. % solids
at a given percent solids is accom-
panied by a maximum in the asso- 72 77 80 83 83 (with oid)
ciated torque curve. Using rheol- 1 1 I
ogy control chemicals at high 900 1000 1500 2000 2500
slurry loadings allows for more Weightof solids,
W,ingrams
torque to be put into the mill. In
large scale continuous tests, the
Fia. 5a—Net Production of taconite versus Ioadina of are in Iabaratorv batch ball mill
total energy input was found to be 4-

as a function of weight percent solids with constan; experimental conditions, including

relatively constant with or with-
a grind time of 30 minutes starting with same feed size.
out rheology control chemicals.
Thus, in the regions of high
throughput in continuous mills
with chemical, less energy is re- assume a consistency that pours mill operating on copper slurry.
quired per unit of throughput to a uniformly and easily from one The rod mill was a 2.5x 4.25 m (8.2
given size. Operating in region C beaker to the other with essen- x 13.9 ft.) overflow discharge mill
causes a drop in total energy draw. tially no solid residual left in the operating at 7W0 of critical speed
Even though it takes a proper original beaker. When this pseu- on an average feed rate of 100 t/h
viscometer to quantitatively fol- doplastic character is exhibited, (110 stph), a rod loading of36% (75
low the slurry rheology transfor- optimal grinding occurs in tum- mm, or 3 in., makeup rod size), and
mation from dilatant to pseudo- bling media mills. Increasing per- a feed size of about 10WO<2.5 cm
plastic to pseudoplastic with yield cent solids produces a situation (1 in.), 50%< 3.35 mm (6 mesh US),
as the percent solids is increased, where the slurry begins to set up, 20V0 < 500 pm (35 mesh US), and
it is possible to qualitatively fol- and it becomes increasingly lWO <75 Km (200 mesh US). Spe-
low this transformation by a sim- difficult to pour any of the slurry cial arrangements were made to
ple laboratory experiment. This from one beaker to another unless keep operating conditions, in-
experiment requires two 1-L glass the sides of the beaker are cluding ore character, as constant
beakers and a supply of constant slapped. It is this yield rheology as possible during the tests.
particle size material having an character that is associated with Changing the percent solids by
upper size of around 1000 pm (18 slower, non first order grinding. changing water was the major
mesh US) and a Schuhmann slope In demonstrating this rheology operating variable. The total en-
of approximately 1 on a plot of log concept to experienced plant op- e$rgy draw of the mill was essen-
fraction less than size versus log erators, a few of them recognized tially constant during all of the
size. A series of slurries are made the pseudoplastic slurry charac- runs in slurry regions A, B, and B’.
up in one beaker by the addition of ter as having the right “feel and Only at very high slurry densities
an appropriate amount of water look’ to maximize grinding. They (region C) was a drop in energy
starting with a low percent solids did this based on years of experi- draw noted. Theological charac-
slurry. With each slurry, the ex- ence without any awareness of the terizations also were performed
perimenter quickly pours the relationship between breakage during the tests to identify slurry
slurry from one beaker to the sec- rate and slurry rheology, pointing rheology behavior at any desired
ond beaker and back to the origi- out the value of experienced plant point in the test program to be
nal and so on for a minute. operators. able to correlate plant results
The dilatant slurry character at with the laboratory data. Four ba-
low percent solids will be obvious sic factors were studied including
from the settling that occurs in the Controlled Rheology Tests in variation in percent solids, the ad-
beakers and the difficulty of com- Industrial Scale Open Circuit dition of chemical GA4272, the
pletely removing all the solids Operations addition of fines to feed, and the
from one beaker to the second influence of slurry temperature.
beaker regardless of how quickly This section presents the results Figure 6 summarizes typical
the transfer is made. As percent of a series of controlled rheology data collected on the basis of net
solids is increased, the slurry will tests run on an open-circuit rod production < 500pm (35 mesh US)
22 JANUARY 1983 MINING ENGINEERING
 n
640 -.
o
0
580 560- ● 560

480 w= 5oog 480- - W=looog 480 w = 1500g

1

‘“u 72 77 80 83
4oo- -
,
72
1
77
I
80
I
83
‘“~ 72 77 80 83

640> 1
I ‘1 o
560 0

a
1A
480 480

“0- “0=
w=2500g
400

72 77 80 83 72 77 80 83 72 77 80 83

Fig. 5b-Net production of taconite in grams versus weight percent solids as a function of solids loading in mill without aid, ●, and
with GA-4272, 0.

versus percent solids by weight. the pseudoplastic region B—now Rheology Tests in Closed
The influence of percent solids occurs at a lower percent solids. Circuit Operations
with and without chemical on the While there is little throughput
normal feed and feed slurry tem- advantage to adding fines to the Using the same copper ore as in
perature of 22°C (71.6”F) is cer- feed in this test, the addition of Table 1 (December ME, P.1669), a
tainly similar in net production fines to rod mills is an important series of tests were run on a ball
and viscosity response to that of concept. This is true because mill operating in closed circuit
Fig. 1 (December ME, issue, pp. those rod mills operating on with a single 610 mm (24 in.)
1669). This is an exampk where coarse feeds that cannot normally hydrocyclone. The mill was an
simple optimization of water use reach theological region B by overflow discharge type? 3 x 5 m
can lead to significant increases water control alone can be forced (10 x 16 ft). It was operating at 80%.
in plant throughput. In similar into this region—hence giving of critical speed on an average
tests run on rod mills at other higher net throughputs—by the in- feed rate of 200 tih (220 stph), a ball
plants, however, the feed size was troduction of fines from some loading of 38% (5 and 2.5 cm, or 2
of’ten so coarse that it was not pos- other point in the overall process. and 1 in., makeup ball sizes), and a
sible to achieve the pseudoplastic Several tests at various plants feed size of about 100%< 3.35 mm
slurry character of region B. This were run to verify this important (6 mesh US), 50% <500 Wm (35
was primarily because the per- concept. The implications of the mesh US), and 20% < 75 Km (200
cent solids simply could not be influence of fines on rod mill mesh US). As in the rod mill test,
raised high enough with normal performance with regard to the energy draw of the mill was es-
water control procedures to safely crusher operating guidelines is sentially constant during all runs
operate in region B. In this envi- obvious. Crushing operations that except in region C, which was
ronment the rod mill will operate produce few fines can be quite lower.
only in region A and optimizing detrimental to optimal rod mill As in the previous section, spe-
water control will produce at performance. cial care was taken to maintain
best only marginal increases in Figure 6 also illustrates the consistent operating conditions
throughput. As explained earlier, influence of a lower feed slurry and theological characterizations
the use of a viscosity control chem- temperature of 8°C (46.4”F) on net were run. The basic factors stud-
ical in region A is not warranted. production. It is obvious that net ied in the closed circuit tests
To more fully demonstrate production is lower. Equally im- were variation in percent solids,
theological influence, the same portant from a control viewpoint, the addition of chemical GA4272,
rod mill was then operated with is that the value of percent solids, the addition of fines to feed, and
lMo of the feed rate being supple- in which region A is transformed the influence of classifier effi-
mented by –38pm (400 mesh US) to region B, is lower than normal ciency.
ore. As can be seen in Fig. 6, a sim- operation. Similar results to Fig. 6 Figure 7 summarizes typical
ilar shape of net production curve also were generated on rod mills data collected on the basis of net
results as in normal operation, but grinding coal and open-circuit production of the total circuit< 75
the maximum throughput—hence ball mills grinding ores. Mm (200 mesh US) versus percent
MINING ENGINEERING JANUARY 19S3 23
I

● Normal feed size and feed slurry temperature of 22°C

■ Normal operation with ().26 kg GA-4272 per tonne of feed ore
40 —
@ . Operation with 10% of feed rate supplemented by minus 400
~ U.S. mesh ore
E o Normal operation with feed slurry temperature of 8°C
~-
0=
:2 35 —
m.
~al
~Q
P P
:: P
.1-
0.2 30 —D
.5 g
~= D
xc
v .-
0
&
z
z
% Solids by Weight

20-
75 76 77 78 79 80 81 82 83 84

Fia. 6-Net production of copper ore qround to less than 500 wm (35 mesh, US) in an industrial scale rod mill as a function of percent
so~ds and slurry rheology type (dila{ant indicated by D, pseudaplastic by P, and pseudoplastic with yield by Y).

solids by weight. As in the rod mill ing circuits. tions of high cyclone feed rate
tests, the influence of percent sol- One of the unexpected results of and/or high cyclone feed percent
id: with and without chemical on closed circuit testing was the solids).
the normal feed is similar in net large theological influence of After detailed testing it was
production and viscosity response changes in classifier efficiency found that the classifier in closed
to that of Fig. 1. In all of the sec- on mill grinding rates. As percent circuit operations can—and does
ondary grinding closed circuit solids were being varied under in many circuits—act as a theolog-
tests run on various ores and coals constant feed solids, it became ob- ical control device on the grinding
at various plants, it was generally vious in many of the tests that the mill itself and that it can be more
not difficult to transform slurry classifier (either hydrocyclone or important than control Of makeup
rheology from dilatant to pseudo- screen type) size separation char- water to the overall circuit. The
plastic to pseudoplastic with yield acteristics were changing with classifier can change mill rheol-
merely by adjusting water control. slurry rheology changes. The ogy by changing the particle size
This control can be done by in- change of hydrocyclone perform- being recirculated back to the
creasing the percent solids either ance with percent solids variation mill and by influencing the cir-
by increasing solids feed rate and has been documented previously. cuit water balance. For example,
holding water addition rate con- In normal hydrocyclone opera- if the classifier (for whatever rea-
stafit, andlor by decreasing water tions it is also well known that an son) presents too coarse a recircu-
feed rate holding solids feed rate increase of flow rate to hydrocy- lated product (and/or with too
constant. The first method is most clones will cause a decrease in the much water), it can become diffi-
useful for increasing the circuit d,, value of separation. The influ- cult to move the mill rheology out
solids throughput while still pro- ence of viscosity control chemi- of region A (dilatancy) even with
ducing a relatively constant frac- cals, added to the makeup circuit large decreases in water addition
tion of solids less than some spec- feed, on classifier performance to the makeup feed. In this envi-
ified size. The second method is was also quite noticeable. Figure ronment, the net circuit through-
better suited for grinding a given 8 gives a Tromp curve example of put is relatively insensitive to
solids feed rate finer than nor- this effect using GA-4272 on a operating changes and is not oper-
mal. Also, in closed secondary stand-alone hydrocyclone oper- ating in a maximum throughput
grinding circuit operations, the ating on a copper ore. In general, condition. On the other hand, if
addition of fines to the makeup it was found that the use of viscos- the classifier (again, for whatever
feed was not generally advanta- ity control chemicals on hydrocy- reason) presents too fine a
geous and thus it is not included in cloning slightly increased the ap- recirculated product (and/or with
Fig. 7. The exceptions to this ob- parent solids bypass fraction, and too little water), the mill rheology
servation were primary closed cir- made the sharpness of separation will have a tendency to be in re-
cuit ball mills and/or autogenous (Sharpness Index S.1. = d,,/d,,) gion C (slurry yield value present).
circuits operating on relatively more ideal (higher S.1. value) This also is not an optimal opera-
coarser feeds than normally asso- while maintaining or increasing tion from an overall circuit
ciated with secondary ore grind- the d,, value (even under condi- throughput basis, although use of
24 JANUARY 1983 MINING ENGINEERING
 circuit throughput depended on
an organized plant testing pro-
gram. This is unfortunate, since
laboratory tests were capable of
I Claaaifier Setting
predicting rheology conversion in
a dm S.1.
150 1- — open-circuit mills but the addi-
● Normal oparation 0.25 185 0.50 tion of the classifier makes such

1
Normal operation with 0.24 kg 0.27 190 0.56 predictions too unreliable. This
■ GA-4272 per tonne of feed ore
points out the necessity of devel-
120 . Normal operation with deliberate 0,26 176 0.49
classifier change
oping a better understanding of
classifier (especially hydrocy-
clone) operation in general and
90 — P rheology effects in classification,
more specifically.

D
80 — Conclusions
This article summarizes slurry
rheology influence on the grind-
70 — ing throughput of batch labora-
Y tory and continuous plant-scale
mills operating in open and closed
circuit configurations. The re-
‘-71 72 73 74 75 76 77 78 79 80 81 82 sults of controlled changes in per–
% Solids by Weight cent solids, particle size, slurry
temperature, and viscosity con-
trol chemicals were demonstrated.
An underlying theological ex-
planation was offered that tied the
Fia. 7—Net Production of coDDer ore around to less than 75 urn . (200
. mesh. ,.US) in an type and magnitude of particle
in~ustrial scale ball mill circui; with a ~ 10-mm (24-in. ) hydrocyclone as a function of breakage occurring in the mill
percent solids and slurry rheology type (dilatant indicated by D, pseudoplastic by P, with the corresponding theolog-
and pseudoplastic with yield by Y). ical character that the slurry was
exhibiting at some given set of
viscosity control chemicals will mill circuit was deliberate varia- operating conditions. More spe-
generally help. tion of the hydrocyclone clas- cifically, dilatant slurries show
The maximum throughput ob- sifier size separation efficiency. first order breakage. Pseudoplas-
tainable from any of the closed Figure 7 also gives the net circuit tic slurries also demonstrate first
circuits tested was when the production of –75pm (–200 mesh) order breakage, but at a higher
grinding device itself was operat- material as a function of percent rate. Pseudoplastic slurries with
ing in region B (pseudoplastic solids for three different separa- significant yield values show
slurry). This is an important con- tion efficiencies of the hydrocy- slower non first order breakage.
clusion. Achieving mill operation clone, including the normal oper- It is important to note that all of
in region B often requires a deli- ating case (a = 0.25, d,~ = 185, S.1. = these theological characteriza-
cate balance of water control and 0.50) and the case with chemical (a tions were done on the slurry as it
classifier equipment setting. = 0.27, d,, = 190, S.1. = 0.56). The exited the mill. In reality (espe-
Operating a closed circuit mill in smaller d~~case over normal oper- cially in batch tests), this does not
region B or B’ can typically give ations (a = 0.28, ds, = 176, S.1. = mean that the slurry was exhib-
throughput increases of 4-20% 0.49) was achieved by decreasing iting this theological character
over operating in regions A and C. the vortex finder diameter. The for the whole time period in the
The use of viscosity control chemi- influence of recirculating a finer mill. It is safe to say that a dilatant
cals makes region B to B’ larger (in product back to the mill due to slurry exiting the mill was acting
terms of percent solids range, for smaller d~~value is evident in the as a dilatant slurry during its time
example) and the effect of region earlier occurrence of pseudoplas- in the mill. Also it is important to
B to B’ more pronounced (higher tic slurry character in terms of note that it is much easier to con-
throughput of region B’ over re- percent solids. This type of test trol and maintain the desired
gion B by 2-127.). It also was found was repeated on a variety of rheology character in continuous
that maintaining a mill in region B closed circuit grinding opera- mills than in batch tests because
may not be easy on a continuous tions. It generally was found that less dramatic changes in particle
operating plant schedule due to the mechanical setting and opera- size and surface area are in-
unavailability of appropriate tion of the classifier (as reflected volved. The recent reporting of ac-
plant online viscometers and, in by changes in a, d,,, and S.1.) will celerated breakage in small batch
some plants, large variability in have a major impact on where mills of larger particles in the pres-
ore type, feed size, and water sup- (from a percent solids viewpoint, ence of fines is a direct conse-
ply. Even in this latter environ- for example) the conversion from quence of the theological trans-
ment, however, if the circuit can a dilatant to pseudoplastic to a formation from dilatant to pseudo-
be operated at least part of the significant yield slurry will take plastic as grinding occurs.
time with the mill in region B or place. However, in most of the cir- In addition to the above theolog-
B:, the increase in net throughput cuit test runs it was not possible to ical characterization is the need,
WI1lwarrant the extra effort. predict this rheology conversion a from a maximum throughput ba-
Another type of test sequence priori, and thus the corresponding sis, of tumbling media mills to op-
run on the same closed circuit ball determination of maximum net erate on as thick a slurry basis as
MINING ENGINEERING JANUARY 1983 25
 1.0 I
Experimental (average of 2 tests)
9 Without dispersant
. With dispersant
.8 —
— Without dispersant 9 I
*
.-
_ — With dispersant
! Feed size
28.5% <75 microns
.
8 48,2% <212 microns
= .6 —

s
>
.-+ 1
Water split without
.4 functions
_ :/ dispersant

.2 s — 0.28 175 0.60

Water split with --- 0.33 191 0.68
dispersant
(~ two ~andard deviations)

0.0
20 50 100 200 500 1000
Size, pm

Fig. 8—Size separation efficiency ofa610-mm (24-in.) stand-alone hydrocyclone operating on a copper ore with and without can.
trot chemical GA-4272. Selectivity value Si is def ined as weight fraction of feed size i to classifier sent to underflaw or coarse stream.

possible that still offers a low of viscosity control chemicals, and ing of slurries, which is desirable
enough viscosity to keep grinding the role of classification were dis- because it allows even higher den-
first order. The above maximiza- cussed. A general observation on sity slurries to be achieved. How-
tion also requires a solids loading both open and closed circuit oper- ever, in region C, the overall net
at equal to or above the media ation was that it was necessary to production also drops due to the
void volume loading. The study have the slurry in the grinding occurrence of non first order
shows that increased slurry den- mill itself be in a pseudoplastic breakage. Thus, the optimization
sity allows for increased solids rheology region to achieve maxi- problem of balancing packing
loading before passing through mum net throughput. Typical in- efficiency versus net production.
the maximum in the net produc- teractions of grinding mill rheol-
tion curve where the fall off is due ogy and classifier operation also
to non first order breakage. It is were presented. The use of the
concepts of this article on indus- References
also evident that use of rheology
control chemicals allows even trial scale have proven valuable to Austin, L., 1971, Powder Technology, Vol. 5, pp. 1-17.

higher net production at even the operating mineral/coal pro- Austin, L., Klimpel, R., and Luckie, P., 1983, The Process
Engineering of Size Reduction, AlME.
higher solids loading. This is ac- cessing firms involved in the Bird, R., Steward, W., and Lightfoot,E., 1960, Transpofl
complished by maintaining first overall testing program. Phenomena, Wiley, New York.
order breakage in a solids loading With regard to the grinding of Katzer, M., Klimpel, R., and Sewall, J., 1981, “Example of
and percent solids regime that dense coal water slurries, the use the Laboratory Characterization of Grinding Aids in tha
Wet Grinding of Ores,” Mining Engineering, Vol. 33, No.
normally would be experiencing of theological control chemicals 10, pp. 1471-1476.
non first order breakage. These is particularly important because Klimpel, R., 1982, Powder Technology, Vol. 31, pp.
trends seem to indicate that to of the penalties associated with 255-262.

achieve the maximum net produc- excessive water in the slurry end Klimpel, R., 1962, Powder Technology, Vol. 32, pp.
261-277.
tion possible in a tumbling media uses. The preparation of such
Klimpel, R., 1960, Proceedings: Symposium on Fine
mill requires that the bed volume slurries is somewhat different Particles Processing, Al ME, pp. 112%1 152.
be expanded as much as possible than maximizing the net grinding Klimpel, R., and Austin, L., 1962, Powder Technology,
with excess slurry until the media production of minerals (for exam- Vol. 31, pp. 23S-253.
material contact changes charac- ple, operating in regions B and B’). Klimpel, R., and Manfroy, W., 1978, /ndustria/ & Engi-
ter. In dense coal water slurry grind- neering Chemistry, Vol. 17,
pp. 516-523.

A series of industrial scale ing, an economic/technicaI opti- Lynch, A., 1977, Mineral
Elseviar Scientific.
Crushing and Grinding Circuits,

operating guidelines are pre- mization problem exists. As noted Manfroy, W., snd Klimpel, R., US Patents 4,126,276;
sented that have proven useful in in Part 1, the grinding of dense 4,126,277; 4,126,278; 4,136,630; 4,162,044; 4,162,045;
increasing the net production ca- slurries in region C causes the pri- 4,274,599.

pability of grinding circuits. Spe- mary breakage distribution B,,, to Mular, A., and Malghan, S., 1982, Design and Installation
of Comminution Circuits, Al ME.
cifically, the influence of per- become finer. Experience has
Wasp, E., Kenny, J., and Gandhi, R., 1977, Solid-Liquid
cent solids, adding fines, de- shown that this finer Bl,j is useful F/ow, Trans Tech Publications, Clausthal-Zellarfald,
creasing temperature, the use in producing more efficient pack- Germany.

26 JANUARY 1983 MINING ENGINEERING