Documente Academic
Documente Profesional
Documente Cultură
LIBRARY
ADDITIONS
Government Publications:
CYPRUS, GEOLOGICAL SURVEY DEPARTMENT. Annual report ... for the year
FRIEDENSBt:RG,
Ferdinand,
und
KRENGEL, Rolf. Die wirtschaftliche
Bedeutung des Metallerzbergbaus und
Metal{huttenwesens in der Bundesrepubhk Deutschland. Berlin: Duncker
U?d Humblot, 1962. 239 p., tabs.
b1bho. DM 28.60.
'
*FRANCE,
BUREAU
of mines in India.
Proceedings of Sympositun held at
the Central Mining Research Station
Dhanbad, 9th to 12th December'
19?1. Calcutta: Books & Journal;
P!IVate Ltd., 1962. 668 p., illus.,
d1agrs., tabs., biblios. Rs. 45.00 or
70s. (Journal of Mines Metals &
Fuels, special issue, 1962.)
ROYAL
= (1oo ~ c).
(100c-])
f)
----r--f) .cio0::-7)
x loo
(c
(c
which transforms to
Rj
(c
f)
. ooo-=~ 7J x 1oo
or to
E
NATIONALCHEMICALLABORATORY.
TION
handbook.
DIRECTION
DES
MINES,
DE
DOCUMENTATION
September 1962
MINISTRY
OF
Mrl'>.'ES
AND
ONTARIO,
DEPARTMENT
OF
MINEs.
!)f)'
where Cmax is the assay grade of a 100 per cent pure concentrate.
This basic efficiency formula is logically derived, is applicable to all grades of
feed and gives consistent assessments of any one operation, whether calculated on
the basis of the concentration of values or of gangue.
By contrast, the limitations of the formula for expressing concentration efficiency
referred to by Fle~ing, and dealt with in more detail by Stevens and Collins, are
shown, particularly"'!n its application to high-grade feeds.
Examples are given demonstrating the use of the basic efficiency formula in
sizing and dewatering operations.
\V'ESTERN
AUSTRALIA,
GOVERNJ'v!ENT
CHEMICAL LABORATORIES. Report, ..
Yy
....
698
Be
le
_jc ~) x R
(emax -f)
and
~ -
1)
R
100
699
E. DOUGLAS:
. For t~e purpose of this analysis ~he concentrating operation is considered m tw.o stage.s, a zero-efficiency sampling operation giving a
percentage weight spht of C and 100 - C, followed by the simultaneous
movement of the values from the tailing fraction into the concentrate and
the gangue from the C units of concentrate into the tailings. The move:Uent
of values governs the final recovery, while both this and the movement of
the gangue influence the concentrate grade to a similar degree.
Both of these ~ransfer .operations can be expressed in simple terms.
The extent to which each IS effected can be related to its maximum value
thereby indicating its efficiency, the overall efficiency being the produc~
of the two individual operation efficiencies.
Normally, concentration is assessed either by chemical assay or by
reference t? the percentage content of a particular mineral. In this paper
the latter IS used as the initial basis for deriving the efficiency of concentration, while in a later section modifications which account for its
conversion to chemical assay are indicated.
The block diagram shown in Fig. 1 illustrates the three fractions to be
considered; the feed (ABDE), the concentrate (AFGE), and the tailing
(FBDG). It c~n be seen that the concentrate fraction has been up-graded
from f to c, with the resultant depletion of tailings from f to t.
A
c=:::J
t.loluob/~
GANGUE:
VALUE:S IN HE:AD
VALUE:S IN CONCENTRATE
minflrol
contoinfld in
of
concrntroffl
Wflight.)
Wrzight of concflntrotfl os a
przrcvntogfl of fflfld Wflight.
1~~)
100
1{m
(1)
100
1~jl
Product
(10~0~ n
I Constituents
(c- f)
(lOO___:./)
100
c7), c and f;
Head
(2)
84
__
1_
100
Grade
%A [
~~
119
I 95 I 5
(3)
20
100
S~6
Recovery
B o;,,
75
988
I 95
I
126
874
1000
M
~-n ~-n
Efficiency of concentratiOn = ( 100c -=-:]) . - - . (Too ~/) x 100
Concentration of ""
Concentration of B
(into A tail)
Efficiency
(Basic efficiency
formula)
Efficiency
(Stevens and Collins
formula)
=
(loo-=- f) R
169
665
c10o
(c - f )
(c -f)
For low-:grade feeds the two formulae are in close agreement; this is
borne out m al~ the examples quoted in the Stevens and Collins paper.
However, for high-grade feeds there is considerable divergence between
the two formulae, 66 5 per cent efficiency resulting from the 'Be' criterion
compared with 16 per cent of the basic efficiency formula. Further considerati?n is necessary in ?rder to establish which of the two efficiency
figures IS more representative of the actual concentration achieved.
A sampling operation, having the same weight split as that of concentrate 'B' in Table I, will be associated with a 95 per cent grade, an 84 per
cent recovery and a zero efficiency.
. In the 'B' concen~rate, recovery has increased from 84 to 87 4 per cent,
I.e. 34 per cent_umts out of a possible 16 (2? per cent), and the resulting
grade shows an mcrease of 3 8 out of a possible 5 per cent units (enrichmen_t ratio 1 0~). Such a result does not seem to warrant a 66 5 per cent
e_fficienc~, par~Icula~ly when compared with the 16 per cent 'A' concentratiOn efficiency m which the recovery, over a simple sampling operation, was
so that
.
(R - C)
(c -f)
Efficiency = (100 _ C). (100 _f) X 100
Ec
=~ 7,
A%
__ l_ _80-~
(c- f)
I
. Recovery Grade
Init~ally, the results in_ Table I can be assessed from two aspects; (a) the
efficiency of concentratmg A from a relatively low-grade feed, i.e. C = 16,
c.= 25, f = 5, R .= 80; and (b) the efficiency of concentrating B from a
high-grade feed, I.e. C = 84, c = 98 8, f = 95, R == 87 4. These give
the efficiencies shown in Table II.
.
I
When the concentrate weight, C, its percentage valuable mineral content, c, and the percentage mineral in the feed, f, are known, the formula
gives an exact account of the efficiency of concentration. It is common
practice, however, for concentrating operations to be expressed in terms of
R (where R =
A tail/ B cone.
.
.
.. Efficiency of gangue transfer
701
E. DOUGLAS
700
(4)
Chemical assays
When products are assessed by chemical assay, c and/ are used to refer
to the percentage assay figures. The formula requires only a minor
modification, namely that the ( 100 -f) denominator of the last term be
replaced by (cmax -f), where cmax represents the chemical assay which
would result from a 100 per cent grade concentrate. In this case:
703
E. DOUGLAS:
increased by 64 per cent units out of a possible 84, with a grade increase
from 5 to 25 per cent (enrichment ratio 5).
It is estimated that a 66 5 per cent efficiency from the low-grade feed
concentration would require a concentrate grade of 661 (enrichment
ratio 13) coupled with a recovery of88 per cent (an increase of72 per cent
values out of a possible 84 per cent). This is considerably more impressive
than the 'B' concentration operation for which the lower efficiency of
16 per cent must be considered more realistic. Also, it is in exact agreement
with the calculated efficiency of the same operation based on the concentration of 'A' from 'B'.
Finallv, with the basic efficiency formula it is possible to operate on a
'concentrate' which has been depleted in values and still achieve a consistent result-for example, 'A' tail (Table I) is considered to be a
concentrate depleted in values, where, c '= 1 19, I
5, C 84, R
20.
Efficiency, according to the basic efficiency formula, is 16 per. cent,
indicating that values have been rejected, or gangue concentrated, with an
efficiency which is in complete agreement with the two alternative assessments of the same operation. The formula discussed by Stevens and
Collins, however, would produce a third and different efficiency, of -08,
for the same operation.
The difference between the two formulae lies in the use of recovery in
the 'Be' derivation of efficiency, as compared with the 'value concentration'
-I), m
t he b asiC
effi ciency
702
REFERENCES
1. FLEMING, M. G. Selectivity factors in flotation. Chem. &
1959, 123Q-8.
Ind., Oct. 3,
Wt.
%feed
24
97 . 6
1000
-241'Wt.(%)
938
2 .6
48
+24~t
Wt.(%)
62
97 . 4
952
4 8 per cent of the feed was smaller than the required 24-p. size limit;
2 ~ per cent of the underflow, which comprised 97 6 per cent of the feed
weight, was finer than 24 JL Considering the +24-JL material as the concentrate, C = 9,_ 6, c 97 4, I
95 2, from which E
43 per cent.
Cyclone: classification
T~e result? used in this particular example (shown in Table IV) were
obtamed durmg a cyclone classifying operation in which it was necessary
to make a size cut at 72 mesh.
_,._ __
xR
200
500
~:,
704
E.
DOUGLAS
SYNOPSIS
Sieving
The results shown in Table V indicate the effectiveness of an industrial
lOO-mesh screening operation.
TABLE V.-Industrial sieving using 1 00-mesh B.S. sieve
Wt.
-100
Product
'J~
Wt. (<)~)
Oversize
15
9 3
Undersize
85
966
Feed .
100
835
,,
This paper describes, in condensed form, part of a research programme into the
dynamic behaviour of rocks. Thin specimens were inserted within a split Hopkinson
bar and subjected to high-intensity stress pulses by means of explosive detonators.
The stress-strain characteristics, derived over a period of about 20 microseconds,
showed that rocks develop considerable hysteresis and have little tendency to recover
over the period of the pulse. This apparent visco-ebstic behaviour would seem to
require for its explanation consideration of a complex phenomenological model
defining a spectrum of relaxation times while, in the case of a porous rock, an air
dashpot should be used to simulate pore deformation.
Mention is made of further investigations which seem to suggest that rocks
selectively attenuate frequencies both higher and lower than those within a critical
band.
IT IS USUALLY FOUND THAT, for very short periods of time, rocks can withstand stresses greatly in excess of their 'static' compressive and tensile
strengths. This phenomenon is sometimes explained on the basis of an
essentially visco-elastic approach whereby a progressive time-dependent
deformation, occurring under constant load, induces failure at some time
later than that at which the load was applied. The creep properties of
polycrystalline materials are not divorced from those mechanisms which
promote an out-of-phase relationship between stress and strain at high
rates of loading and in both cases it becomes necessary to consider the
contribution, both in percentage and disposition, of the several minerals
comprising the rock towards the deformation of the rock as a whole.
Deformation p~cesses in a time-dependent body can be visualized in
terms of phenomenological models whereby spring and dashpot systems in
series or parallel idealize elastic and visco-elastic behaviour within that
body. The springs deform under Hookean conditions, their stiffness
defining the appropriate Young's Modulus while the viscous dashpot
elements obey Newton's law of viscosity. The Maxwell model (a spring and
dashpot in series) is the simplest and has been used to describe the response
of high polymers to stress. It greatly idealizes the behaviour of a real solid,
however, for in general there is some delayed recovery in most materials.
The Kelvin-Voigt model exhibits viscously controlled deformation in both
the loading and unloading modes (spring and dashpot in parallel) allowing
for no purely elastic strain. A three-element model-the standard linear
solid comprising another spring in series with a Kelvin-Voigt model
*Paper received by the Institution of Mining and Metallurgy on 1st March, 1962,
and published on 6th September, 1962; for discussion at a General Meeting on
18th October, 1962.
tPostgraduate School in Mining, University of Sheffield.
705
281
DISCUSSIONSAND CONTRIBUTIONS
' loo'
iioo c17
B^src EFFrcrENcyFoRMULA_DlscussroN
E. DOUGLAS: DERIVATION OF A
282
thcn
-r
-(,' .L).9._
; too.
100
J)c
the individual eliciencies of the t!'!'o
possible
to
indicatc
Thus it was
oDerationswhiih iDfluenced the conccDtration proccss' They could be
cimbincd in a number of ways,eachof rvhich would adequatclyrcprcsent
thc olcrall eliciency. For eximllc. it could be cxprct'eda' the.arirhmetic
-."" ot rl.t" geomeiric m"an, & et en morc simply as the product of the
i*no t.t-t. l-i" ptoduct combinationhad been uscdin deri!ing the formula
paper.
'orcsentedin the
Scveral variationi could hc applied to the generaltheme used in that
dcrivation; for examplc, a sliglitly more logical, but considerably more
293
Ca /i1
! t00
2C(t
fa)
C.!)
Il
: Tqtoo cl
Cb . rb
/,,100
tb,
cb;
( f ,..,
li,
tuu
7 3 -r o o L . , 1 '
r^^
ruu
(l)
_ {r.
KOOrn50n.
o f o )+ c b ( c b - t b ) , C g ( . s - r s )
t = C o _ t ._
x/oo
( / O O - C a ) t a + ( O O - C b ) f b + ( / O O- C g ) t s
C (c 'f )
--
>: f (/oo,c)
xtvv
FiE..4.
complicatcd,rcpresentationtould be had by introducing a factor expressing
ihe iate nf raiuc" conccntratio $ilh r('pcct to ganguereicctiolr' An
alternativebut almostidenticaltreatmentcould bc appJiedto producc.an
efrciency formula to givc directly rhe overall efliclencl' ot an operatron
from which sevcral conccntratesivcre produced. One o[ the two block
diagramrin Fig. A abore rcpre.cntcdthc fccd anJ the olhcr thc product'
to be
ui inna"ntrrtio"n..A\.uminAth( eiicicner of collcctltralloD
"
284
OF A
E. DOUGLAS;
DERI\IATION
BAsrcEFFIcIENcyFoRMtTLA_DrscussIoN
295
[, DoucLAs:DERtvAT]oNoF A
286
Eiicicncy =
usirutwonr"*-
T1ASS: POTENTIAL
DIFFERENCE
luass >.PofEN-iAi
DIFFERENCE).,A"
/r).
r l c c i m aflr a c t i o n )
,'ra'
..^tatI)
o r a l t c r n a r r \ c l1) ( .
{lO0
,)
Lc
pcr cent.
t'
/'l
q;
(l
h)lh
'1[Valucs].r:l(Massof Values)
Thcrcforc'ba.icefficicnc]',,,;
-, ' alue.l".* . / (Ma:. qf Vslus\)', .''
There was surely no analoguefor the author's form ofcfficiency expression
and he suggestedthat thc'basic cmciency' lbrmula was empirical while the
'fundamental'. Morc seriously
formula for concentrationemciency(Er) was
still rhc denominator of thc basic cfficiency formula containcd a variable;
that was untcnable,The denominator had dillerent valuesfor separations,
of the same mixture, that rcsulted in equal grades of concentratesbut
dillering concentrateweights. Thus it was implied that the maximum
uscful work that could bc done in completcly separatinga gi\'n mixtur
was a variable. He lclt that -\'Lr.Douglas should look a littlc more closely
at the implicit assumption in the bald statemcnt given for efliciency of
transfcr of valucs on pagc 700, linc 4.
For all the cxamplcs given in rhe paper, the formula for concentration
cfficiency,Ec, gave the cor(cct cvaluationin terms of u:eful work achicved
comparcd to the maximum possibleuseful work.
287
Ba.icetfitien*[-
c
R,,,,
Jb
.71',
",
loo c '
( l0o Ci
E . D O U G L A S :D I R J \ ' i T I O N O F A
288
t00
o
+
x
o
lore,oqe)
,-x-^-x
>_a;
it io
Q=
IQ
25
t00
72
52
36
200
50
100
S I Z E O F S E P A R A T I OSN ( 8 . S .m e ! h ) lo9 p sccle)
REOUIRED
Fis. B.
Re/atinshit
b.tlten
a claxillcat"'t
289
5.9
91.1
r000
32 2
0 16
20
6? I
9q 8.1
980
9i O
5 0
1 0 00
4.1
95.9
l o ( t0
29.3
91. I
28.6.
[. DoucLAs:DERTVATIoN
oF A
BAsrc EFFrcrENcyFoRMULA-DIscusstoN
290
291
292
T . D O U G L A S :D F R I V A I T O N O F A
293
"" ( R C
) ' f c)c 100 J
(100
Thus for thc same recovery (R) and rutio of conccntation (1./C), the
/
elliciency indcx waspropo.tional to
,o gir rng a highcr eHiciency
tOO i,
index to a higher grade feed. This upp""."'d to be illogical.
(4)^lf a plantwerc opcraredat lughiecoveryand lori ratio oI
concenrra.
tron lor onc pcriod, and then low recovcryand high ratio of conccntratron
for anouler
anothcr^period,
ror
pe!rod, tJlc
rlrc cthclcncy
cllicicncy tndex
index lor
piriod rnight be the same.
foi each
eachpcriod
trowever, ifthc
However,
lr rc two
penods rveie
t\\.o periods
$'ere taken
taken togcther
togcther a very
verv differcnt elliciency
ejliciencv
index migbt be obtaincd. Thar also appeired illogical, an.l coutd leal iir
difficultics in intcrpreting results.
Tltat urasbest iliusrrarcd [.v the follorving cxamplc:
Irt.
Fccd
Conccnrlilte
Tails
2nd t etiad
Feed
Concntrutc
Tails
Fccd
Cor:ccntrlttc
Triis
Grale
rR.foz./y
''.,
100
35
05
15
40
15,1
935
l(10
7.38
92 62
15
80
1l
392
)o0
?1.2
768
t5
17
6.1
Ratio of
.otl.cntr,ttiln
286
EfJiciency
ozLlex
264
lt 5
264
17
2 1. 6
294
E . D O U G L A S :D E R I V A T I O N O F A
Mr. D. N. Moir said that in vicq' of what had bccn said by prcvious
speakersv'ho wishcd to introducc economicsinto the formula, hc fclt that
it was an appropriate moment to defcnd the researchand dcvclopment
mincral engineer against a common criticism rvhich rvas levelled against
him whenevcr he attcmptcd to bring out a formula rvhich rvould enable
him to assesshis test results accuratcly.That criticism usually took the
form:'It is all very well for you chapsto invent fancy formulae to give you
an idea of what you arc doing, but you are neglcctingthe economicsofthc
job; on the plant we have to set grade and recovery according to the
economicsof the proccssand your formula does not hclp us.' He thought
that confusion aroseout of a misunderstandingabout the real purposc of
such a formula,
He was surc thc good researchand developmentcnginccr lt'as quite rvell
aware that his process must be economically sound and also, while
mctallurgicall.va ccrtain rcsult rvashighly desirable,the plant targct might
be different and was adjustedto suit the pafiicular conditionspertaining at
any given plant undcr thc prcvailing market conditions.
At the bench tcsting stage,the rcsearchand developmcntcngineermust
work up his processto the highest peak of performanceto show the vcry
best metallurgical rcsult Ehich could be expected.It rvas at that stage of
the opcration that the formulac lirst madc their appearanceand they did
so in order to cnablc the engineerto assesseach part of thc circuit and
bring it to its peak of performance.
Subsequently,in small or large pilot-scale testing a rangc of rcsults
would be indicatcd and sulicient data collected to cnable a series of
calculations to be made which would indicate the approximatc level of
performanccnccessaryto achievethe correcteconomicstating into account
such factors as nature and size of ore deposit, grade, mining ratc, capital,
transport and crection costs of equipment, local labour costs, rate of
amorrization, etc. (assuming throughout that the process appcared
economicallysound at the start).
Final economics would bc scttlcd $,hcn a detailed design study vas
undertakenlbr thc l'ull-scaleplant. Ilven when thc plant wasruDning, small
changcsin pcrlormance might have to bc madc from time to time to mect
changes,for examplc,in plant input or in the market pricc ofthc product,
the result ofrhich would be that the finallevcl of pcrformancemight well
be diferent from those indicatcd in thc initial studies.
The formulae rvhich had been discusscd,those of Stevensand Colhns
and ofDouglas, and othcn, '"r'crcnot in any u'ay intendedto solvecconomic
problcm' and it \\a\ unrearonable
to expectthcm to do so.
The economic problcms were individual to a particular plant and
orebody and which gcneral formulae could not bc cxpcctcd to solve.
In addition to thcir use in developmentstagcsof the rvork thc application ofthosc formulac in gaugingthe day-to-day pcrformanceofindividual
parts of thc circuit should not bc overlookcd; sornc of those applications
had been discussedbv thc abovc authors.
IiIRITTIN
REMARKS
295
CONTRIBUTION
75
80
85
70
15 0 42.5 40 0 3'7 5
90
32 5
95
25 0
975
15.O
90
51 0
95
41 7
97.5
2t.4
70
600
i5
605
80
605
85
600
325
955
25.0 150
940 900
296
E. DOUGLAS: DERIVATION OF A
(-onc.ntratorrccorcr',
(lincent.rte
srade
,,
'I
sale oi
.
Itnrins,
millins
xnJ
*eneral expenses {
snclrins cos({
t0
15
7!:r.i
I50
77U
:l ,110
1jj
-3l S :
656
r,.r.-
1l3ia
1.00
l0 00
3lllJ
41.
11 ll
.1 lol
J)a
lio
l:l
tio
.t 165
'1,16.5
l5 i)
4 3rl?
113 0
165 6
3lll 0
llo
129 7
1500
1522
raa 0
r:0
1 5 ?5
154
.lI'.2
71 6
E:rr
241 a
26.1 6
271 9
:S7 I
10,!.I
33t :l
llt
12) I
|)4 2
t.t: 6
r12 I
r06 ;r
mrrkcr a
TDIIL .osi ol mebl
ductiond.
s0
l00
211 t9a 3
43A.1
165 0
52 3
Even this does not tell the u,holc storv. If a leaching proccss rvcre
availableas an alternativc to sDrelting,it might well bc profitable to accept
a concentratcgrade of 15 per ccnt in ordcr to achiele the high rccovcry
of97.5 per cenr.This rvoulddependon leachingrccovcryand costs,and
would rcquire a further set of calculations.
Although the costs and recovelies that I havc uscd are hypothetical,
I submit that they are typical of actual practice. Even then, thcy arc an
over-simplification,
since to obtain the higher rccovc|y $ould probably
involve highl-r capital cost of plant, and alloq'ancel ould havc to bc-madc
for rhis in assessingrhc cconomicsof the operation. To the engineer or
manager,statistics, costs and emcicncy figures have a value only if thcy
provide reliable information on which to base dccisions.Anyone basing a
decision on an index rather than on thc economicswould bc dccidine
blind insteadof in thc licht of the facts.
AUTHOR'S REPLY
Mr. E. Douglas: As indicarcd in my presentationl,this papcr was
preparedinitiallf in the form of a discussionon the pape( bv Stevens
and Collins which I had assumed,apparently erroncously,would be read
by membersin conjunction with mr paper. The most common misunderstanding has ariscn from the readers'pcrsonalinterpretation of'elicienry
formula for concentrating operations'. In this papcr it refers to the
eliciency v'ith rvhich a mixturc is separated into its individual components. A numbcr of contributors havc rccognized this but others havc
interpreted it as an cconomic eiiciencl'. This lattcr interpretation is not
corrccl; the formula I havc derived caznolbe used lor complete economic
guidancc in relation to thc products from a milling opcration.
My reply to the discussionis in tu o sectionslthe first is concernedwith
REPLY
297
to bcanalosous
torrreproduct
or,h"n'r.r")lt'i,.,iirllilli$l'ilt'"1;';;
oF A
E. DoucLAS:DERTVATIoN
BAsrc EFFrcrENcyroRMUrA-AUTHoR'sR[pLy
recovery term as it stands Mr. Collins is stessing this part of the data
in relation to pcrcentageincreasein gradc. If bv 'minclal dresscr' he is
referring to the plant engincer thcn his argument takesau economicbias.
Altemativel.v,if his referenceto mineral dresserindicates experimenters,
suchas himscl{,workrngon flou'sheetdevclopmentsr
then this is'bcgging
thc argumcnt'.
I maintain that a conccntratingopcratiou is achicvcd by a combination
'l'he
of value concentrationand ganguerejection.
former influencesboth
recovery and grade and, consequentl!',changesin both of thesc factors
should be represcntcdwit}I respect to a sampling operation.
Mr, Collins is in disagrcementwith me and, ashe points out, with other
autho$ he has cited, who consider that the same eficiency of concentration should rcsult whether considcrcd from the aspect of concentrating
values or of concentrating ganguc. He maintains that it is reasonableto
cxpcct two efliciency ratings for any one operation, He usesmy examplc
(Fig. 2 and Table I\r, pp. 702,7O3) to dcrnonsrratchis argument. He
shows (Fig. B), using his expression,thar rwo cut points are produccd,
olle at 100mesh bascdon cyclonc ovcrflow and anothcr at 200 mesh based
on cyclone underflow. Surely, a cyclonc operation has only one cut
point. The basic cfficicncy formula indicates only onc, i.c. 150 mesh,
and this is in exact agreementwith the result produced by the commonly
298
C'(c
/
(i.rxx
f)
-/,
..,rno
'..anrp,l
,l
299
h.rL^,1
,\,tr. Collins, assuming ccrtain intcnrions lor the c)'clo[e products, r.e.
ovcrflow to flotation, underflorv ro tabling, claims that by assessingthe
overflow on the basisof his cfficicnq, formula ir bccomcsa bctrer flotation
fccd than if it were assessedusing rhe basic efticiency formula. A similar
conclusion is reached with rcspcct to thc undcrflow product. I cannot
acccpt the suggestion that the type of expression used in assessinga
product can inllucncc its composition or its rcacdon to subsequent
processcs.
From Mr. Collins's graph, both his efiiciencycharacteristicsarc highu
than the single basic efficiencycharactcristic,and the latter cannot thereforc bc 'a mcasure of the average efficiency of the separation which
incidentally could bc obtaincd by thc addirion of thc tcchnical cfficiency
for both coarseand fine sizes'.
I find it dilicult to understand the comme[ts contained in the first
9 :;J i, 11.r"
"o--6n1y
J
rccognizedco-relation bctwecn rccovcry and ratio of conccntration. Only
/ is constantbut, for any one value of c, the R 1C relationship is a straight
line and for different values of c we have straightJine palamctcrs, all
passingthrough zcro, but rvith clillering slopes.Introducing a bias, eirher
by specifying a constant gradc or a constant rccovcry, considerabll'
simplifies the problem. In this case, as .l\1r. Moncrieff points out, the
cffectivcnessof separation can be expressedin terms of the remaining
variable, i.e, recovery or grade.
My answer to the fust of Mr. Moncrieff's four specific questionsis to
paragraphof Mr. Moncricff's remarks.Surel-v,R
300
E. DouGr-AS:
oF A
DERTYATTON
(i)
(ii)
i;i)
cRcl
35
i.)a
21 2
q3 5
39 2
66
40
80
47
15
15
15
BASICEFFICIENCYFoRIItILA-,1IirHoR'sREpLy
301
obtained by running periods (i) and (ii) for dill'erent times, rvill result rn
furthcl changesin the clliciency ofthc conccntrating blcnding scquence,
Fig. C is prcsentedin order to give a risual illustrarionofihcri arguments and to indicate the characteristicsof thc cllicicncy formula. Hire,
cflicicncy has becn plotted in relation to recovery for vaiious concentrate
grades.In rhir pre.entarion.the conccntrarcgrade detcrmincsthc particular operativc parameter and points r, y and : reprcsent the quotcd
lst, 2nd and combincd period conditions.'l he blendinq curve, which
shorvs thc cfflcienciesro be achieved by various cornbinations of the
l\t and 2nJ pcriod operarior:.has al.o brtn addcd.
E :
c =
coltccht/ole
groda
;az-
-,_/
R7"
):ig. C. F.fliciorclrharauerixics.
I am indebted to -N{r. Noakes for cmphasizing that thc application of
such a formula should bc carried out by opcrators who undiistand such
'tools'.
Also, I am gratel'ul to him for clarifying the dillerences bets,cen
the subjecr matrer of rhis paper and the economlcassessments
of existino
'fhese
milling opcrations.
a'rJtwo totally diffcrcnt factors.l.i;;;;p.ifiiy:
Mr, Noakesconsidersthe latter to be most important ancl I agreeentiiiy
with this- aftitudc. However, thc concentration cfliciency, 6r index, ii
rnost useful at the sragebefore a workable flowsheercxistj. For example,
onemay hate to a"'e\\ lhc comlaraliveamcnabilirie.ofa parliculardcp;.ir
to a numbcr of po..ib)c t:.citmcut merhod., e.g. flotatiou, macncric
sepatatio[ or tabling; ar rhis stage it rs muc-h too earlt to con:idjr the
complicatedand specific economL factors involved at ci ery level of each
302
[.
303
DOUCLAS
of these processes.The obvious necd of a reliable assessmcntof concentration is reflcctcd by the numbcr of papcrs, publishcd by well-known
authors, on this subject.
A reolv to Mr, Dell's contribution will scrvc to demonstratethe considerab-ledifferences betrveen economic and physical assessmentsof a
concentratingopcration. At thc sametime, it will present a generalcomment in answerto Nlr. Heath and others,\vho havc introduced economrcs
into this discussion.
Mr. Dell askswhy economicshave been ruled out. The simple answer
is that I have not attempted to produce an economic asscssment;my sole
concern was to derive an cxpression to indicate the effectivenessr:r'ith
which the constituents(both valuelessfor this exercise)of a mixture have
been scoaratcd. Immcdiatclv suitable economic factors are introduced
the resuiting expressionbec'omesspecific to one operation, at one particular sitc, in one particular country-it is not universal.
Commenting, at Mr. Dcll's request, on his papers, may I refer briefly
to his most recent one 'A comprehensive critcrion of coal-cleaning
efficiency'. Knowing Mr. Dell's experiencein the coal industry I havc
confidence in the adequacy of his assessmentsof coal-cleaningoperations, but I consider thc title misleading. As with combustion efficiency
which relates the heat extracted and the total heat contcnt of the coal.
'coal-cleaning efficiency' suggests the elficiency with which the noncombustibles have been removed, and should thcrcfore not include
monetary considerations.In fact, tlle paper presentsan cconomicaccoi)rt
of a cleaning operation. As an example of this I rcfcr to Mr. Dell's term
(specificprofit) in which the cost of cleaning in the denominator is zero.
This only obtainswhen no cleaning is carried out and thcrcfore, such an
expressiondoes not assessthe cleaning efficiency.
The facts that the expressionsdevelopedby Mr, Dell are indcpendent
of changesin concentrateweight (a constant is rccommended),and that
t}te influence of feed grade is not directly accounted for, demonstratc
clearly the spccific nature of his economic efficiency.A similar degreeof
specificity, I am surc, will apply to almost any other worthwhile economrc
assessment.
Again, if I may be allowed to stressthe point, I agreewith Mr. Dcll,
Mr. Heath and others that an economic standard is of utmost and ol'erriding importance. Horvever, other emcicncy facto$ arc required in the
initial development stagcsof a proccssand, as I have already indicated,
tlis was my objecrive for the basic elliciency formula.