Technical Note:

An innovative technique for minimizingoperational costs of end-user groups ina production environmentby C.J. Alberts*, M.W. von Wielligh*, and G.J. Delportt

Synopsis

The perception exists in many organizations that, once anenergy management system is employed, there need not be anyfurther adjustments to such a system. This is not true, as thefield of energy management is a dynamic field, and constantfeedback and adjustments are needed to enhance the efficiencyof a system to the point that substantial pay-back can beobtained. It is stressed that for any energy management systemto be successful, careful planning should permeate through theorganization, from the top level decision-makers to theworkplace. This paper describes a dynamic technique by whichenergy management strategies can be evaluated by viewing theeffects of the strategies on the organization as a whole.

Introduction

The mines in South Africa are largeconsumers of electricity. In 1985, 27 per centof South Africa's national electrical energysales went to the mining industryl. Thisfigure decreased to 22.6 per cent in 1993.However, the total national demand increasedduring this period. It is evident from thesefigures that least-cost energy planning andenergy management is a necessity in themining sector of South Africa.

Energy management and energymanagement systems:

~ cannot be a one-man show~ cannot be self-maintaining~ require continuous effort, innovative

ideas, and education~ require equipment modifications2.

In the past decade, electricity was sold ona constant rate tariff. Then schemes such astime-of-day metering, peak-demandpenalties, and volume discounts wereintroduced. These schemes recognise that theproduction costs of electricity vary accordingto both long-term and short-term factors.Electricity, as a commodity, is very muchsubject to economic concepts of supply anddemand.

Keywords: energy management. cost function,

real-time-pricing (RTP), systems modelling

The Journal of The South Afncan Institute of Mining and Metallurgy

A more desirable pricing scheme forelectricity would allow for a variable sale pricethat depends on the spot price of power at anyinstant in time. This type of pricing schemehas been given the name Real-time Pricing(RTP). The RTP tariff allows the price ofelectricity to reflect variable production costs.Consumers may elect to modify their use ofelectricity during peak conditions, therebyreducing peak demand and minimizing thecost of electricity, and taking control of usagepatterns and resultant costs3.

In South Africa, as with the rest of theworld, manufacturing industries are goingthrough a period of major of change. Manyorganizations are currently searching forcompetitiveness through business process re-engineering, which focuses on removing non-value-adding activities from the organization.As manufacturing contributes to a highproportion of production costs, systems needto be in place to eliminate waste, continuouslyimprove production processes, and increaseresource utilization4.

New electricity tariffs, like RTP, areenvisaged for the future, hence the need for areal-time energy model. Such a model shouldbe able to read outside inputs, calculate loadprofiles, log historical data, make adjustmentsto certain loads, minimize the operational cost,and interface with various control centres toinfluence the operational decisions.

The aim of this paper is to describe aninnovative technique by which time-varyingenergy tariffs can be implemented in aproduction environment.

Systems modelling

It is stressed that in order to study the energysystem on a production unit effectively, it isnecessary to categorise the production unitinto different sub-systems.

* Iscor Mining, clo Chamber if Mines if SA, MailingDept., Box 6, Iscor.

t CentreJor New Electricity Studies, University ifPretona, Pretoria 0002.

@ The South 4frican Institute if Mining andMetallurgy, 1996. SA ISSN 0038-223X/3. 00 +

0.00. Paper received Aug. 1996; revised paper

received Mar. 1997.

MAY/JUNE 1997 149 ....

INPUTSSub-System

configuration

~Productionprofile

Managementbusiness plan --------

;;Systemconstraints

Supplyconstraints

Productioncosts

11Maintenance

costs

Capital

External /factors

An innovative technique for minimizing operational costs

In order to determine any potential cost saving, forexample a new energy tariff, it is necessary to simulate theproduction unit as a whole with regard to the necessary'What-If scenarios. It is therefore necessary that detailedsystem configurations be compiled for each sub-system.

A building-block concept will be used to model each sub-system. Figure 1 is a representation of a simple modeldeveloped for a subsystem.

This simple model needs to be expanded for eachdifferent sub-system as the requirements differ for eachmodel. The inputs and outputs of a typical sub-system thatneed to be studied is shown in Figure 2.

The development of the cost function will be discussed,as well as its importance to the modelling process.

Cost function

Energy models are powerful tools for studying energyproblems. In practice, the only means of constructing a large-scale model is by the careful integration of separatemathematical descriptions of components of the overallsystem. This mathematical description will be referred to asthe cost function.

The cost function is situated in the process block, asviewed in Figure 1. This cost function will include all the costoptions to be considered in determining the various outputs.

Figure 1-Example of a simple model

Figure 2-lnputs and outputs of typical subsystem

~ 150 MAY/JUNE 1997

The total cost of operating and maintaining theproduction unit can be calculated by adding all the varioussub-system costs together. The total cost of operations of theproduction unit per hour, Cr, is given by the relationship inequation [1]:

k

Cr= ICsnn=!

[1]

where: n = 1,2,..., kk = the number of sub-systems.

The factors Csl ,...,Csn, represent the different costs foreach sub-system in the production unit.

Each sub-system cost, Csn,comprises the different costsas described by equation [2]:

Csn/h = (Cpn + Cmn + fen + Cen),

where n =sub-system numberCpn = production cost

Cmn = maintenance costfen =external factors costCen =capital cost.

The reason for developing a cost function is to supply themanager with a powerful mathematical tool to simulate

various conditions, in order to produce the most economically

sound strategy to manage a plant or a process.

[2}

Coefficient technique

When computing the different cost factors, as described inequation [2], it is necessary to find a correlation betweeninputs and production. For example, there might be a linearrelationship between production and energy consumption.Thus, the cost of energy for a time slot can be computed bymultiplying the production, energy tariff, and an appropriatecoefficient together. The problem at hand is finding this

OUTPUTS

Real cost (R)per time slot

Real cost (R)for different

'what if'scenarios

Realproduction (t)per time slot

Real cost (R)per production

The Journal of The South Afncan Institute of Mining and Metallurgy

An innovative technique for minimizing operational costs

coefficient. The most accurate and also the easiest method bywhich such a coefficient can be found is by using datasampled from the actual production environment.

Equation [3] shows an example of how a coefficient or acost factor can be computed for a cost function.

Cost per hour = (tarift)(production rate)(coefficient)

= (1~)- [T.(k~h)]- [p(*)]- Ce[3]

Where Ce= an energy cost factor.Installed capacity multiplied by kw delivery y production,where production might be a rate of tons per hour or litresper second, for example:

x[kW] ~ y[~]

= Ce - y[~J.x[kW]

x[

kWh]Therefore Ce = y t .

If no linear relationship exists between inputs andproduction, two routes may be taken. Firstly a coefficient canbe computed for each state of production (if relevant), or thenon-linear relationship can be found by the use of curvefitting.

These coefficients must be computed for each input to thecost function. Equation [4] shows an example of a completecost function.

[

(To XPXCpX)+(Tmd XCmd)

]

Cost =+(Cmp X p)+ Cman

where: Cpx = energy coefficientCmd = maximum demand coefficientCmp = maintenance coefficientCman = labour cost coefficient.

This function takes into account energy, as well asmaintenance and labour costs. It should be noted that eachproduction environment will have its own cost function andrelevant inputs.

By producing cost functions for the various subsystemsof a large system, the total cost of production can easily becalculated using computer technology. 'What-if' exercisescan then be done without much effort. This type oftechnique will make the search for the optimal point ofoperations very easy.

[4]

Case Studies

These examples show how the technique can be used foroptimization purposes.

Figure3 describesthe flowof data in the model,showingthe logical flow of data between the input matrix and theoutput matrix. The input matrix represents certain variables,relevant to each time unit, as described by the cost function.These variables form the real-time inputs to the costfunction, which will then produce real-time outputs (theoutput matrix).

The case studies represent a fictitious production plantwith an installed production capacity of 500 Uh. Modelling ofthe plant yields the cost function shown in equation [5].

The Journal of The South Afncan Institute of Mining and Metallurgy

Input sheet Output sheet

Figure 3-Flow of data between matrices

[

(Te XPxO,O4)+(Tmd X2,?)

]Cost per hour = .

+(26xP)+1O6

The input and output matrices will be representedgraphically. Only the variable inputs will be shown by theinput graphs. The output graphs show the operational cost,maintenance cost, and the total production cost.

[5]

Case Study 1

Figure 4 shows the production schedule and energy tariff(Eskom MEGAFLEX)of a typical scheduled production day.As already discussed, the constant inputs will not be shown,but clearly these influence the output graph.

500

450

"'". 360!g300]~250'g 8.200'H,'so..

'00so0

~

20

181614

'2 ~~

-10 .8 "'-6.20

8N 8$!~ 8

oD ~ 8~

8:t

8~

8~

8i<i

8t:I

Time or Doy

I--Productio" proll. -+-Energy tan" I

Figure 4-Case Study 1: variable inputs

The data relevant to the graph is:

Total production for the dayAverage operational cost

Total cost of day's production

Figure 5 shows the output graph relevant to the inputvariables (Figure 4).

1,' ISOO

<50

=.13QO!l

-:H"'" I.100

1.2

1 1

10,8j ..8l] 0,'

0.2

08;; 8N 8. ~ 8.. ~ ~ 8

:ia:!!

8i! I a

1:1TIme .rl>oy

I--Production..., M co..-- Total --><--PlOduc,.." (IIn) I

Figure 5-Case study 1: output graph

MAY/JUNE 1997 151 ....

An innovative technique for minimizing operational costs

Case Study 2

Figure 6 shows an altered production schedule. When thisschedule is compared with that of case study 1, it is clearthat production is increased during the night. This is doneto take advantage of the lower electricity tariffs during thenight.

500450

"""~$:.~~250g boo

~ .§ 150..- 100

500

80

201818,.12 -10 ~i8 .: ~8.20

8N 8

"8oD 8.. 8

$!8ti

a.;!

8~

8~

8j;j a.~

Tlm. of DoV

1--- Produclion profile -+- Enorvv lorif I

Figure &-Case study 2: variable inputs

Figure 7 shows the output graph relevant to the inputvariables (Figure 6). The data relevant to the graph is:

I

Total production for the dayAverage production costTotal cost of day's production

1.8

I."

114

~ 1.2

.. 1j

i:: I] 0.41===1=10.2

1500...,

400 <:-

3501]3Q) i2501 )-.-:-:-:=:=:=:=:::::: , 200 i150....

= 100 !50

0/'~---'......

0 ~ -'-'~-'-'-' '80

8N ~ 8

.. ~ 8$! ~ ~

8\!! ~ a

~a~

Ttm. of aav

1--Pmd~!;~ ,.ot -11- Mat".~~. 000'-- T.'~ '00' Produc'ion(R~) I

Figure 7-Case study 2: output graph

The results show that profit can be increased by shiftingproduction to low-energy-cost time slots. This is not new,but illustrates how easy it is to assess operational costs byusing the developed cost function.

Case Study 3

Just examining the effect of production shifts on energy costalone is short-sighted. The cost function should display theeffect of an alteration on the financial results as a whole.This case study will illustrate how excessive productionabove design capacity can increase maintenance costs,although the energy cost decreases because of better energytariffs. The cost function at hand consists of a two-partmaintenance cost: scheduled maintenance and unplannedmaintenance. Excessive production can cause unplannedmaintenance and downtimes that will have a negative effecton operational costs.

~ 152 MAY/JUNE 1997

600 2D

18,.,.12E i10 -~8 .: ~..20

500

~~400

~~300u ...h200~...

100

0

80

8N

8$!

8:

8"

8oD

8.. 8N 8~

8~

8~

8tj

Tlm. or Doy

l---prod""!;.n!,,.I'. -+-EI18fQv,arifl

Figure 8-Case study 3: variable inputs

Figure 8 shows a production schedule where maximumutilization is made of lower energy tariffs, by drasticallyshifted production schedules.

Figure 9 shows the output graph relevant to the inputvariables (8). It should be apparent that the ratio ofproduction costs to production output worsens whenproduction decreases (see Figure 9, between 07:00 and11:00). One of the main reasons for this, is that utilizationof resources reduces. (A worker needs to be paid, if he isactive or not.) The data relevant to the graph is:

Total production for the dayAverage production cost

Total cost of day's production

Although most of the production takes place duringtimes when electricity is cheap, owing to high maintenancecosts, the overall cost is much higher than that for the twoprevious case studies. The higher maintenance cost iscaused by unplanned maintenance. This happens due to thefact that, in an attempt to produce the maximum whenelectricity is cheap, installed capacity is exceeded andbreakdowns induced. These breakdowns may also be due toscheduled maintenance not being adhered to.

5

..5

. ---L.5

I 3

§ 2.5

i 2

'0 1.5

~ I

800

- 500"

40011

30011200 nlOO

..i~ ... ~. :-:-:-:-:-:-:-:~0

80

~-.~-.~-.------.-.-.-.8N

8

"8oD

8.. 8$!

8ti

a.;!

8~

a.~

a.~

a.~

rima of aav

I-+- Prod",".n '00' --- Mat".no",. 'oa' -- T.tal .0.' -- Produc'lon'R~) I

Figure 9-case study 3: output graph

Interpretation of the results

An evaluation of the relevant data fromthe three case studiesshowsthat case study2 has the most financially soundoutputs. Table 1 shows a comparison of the case study outputs.

The Journal of The South African Institute of Mining and Metallurgy

An innovative technique for minimizing operational costs

Case study 1 is negatively influencedby the fact thatproduction takes place while electricity costs are high.

Case study 3 is negatively influenced by highmaintenance costs due to breakdowns. This is a perfectexample of why the cost function should take all possibleexpenses into account.

Case study 2 is the best option, since the requiredproduction is achieved with the lowest total production cost.

Table I

Comparison of case studies

Casestudy3

85503.06

26123

Summary and conclusions

This article has discussed the development of a compre-hensive cost function that takes into account all the factorsthat influence the total cost of operation.

The function presents the user with a means of quicklyand easily computing finances of production units, with aspreadsheet or by the use of a computer program calculatingreal-time results from real-time inputs.

This technique, will become extremely useful with theimplementation of real-time electricity pricing, in caseswhere production schedules differs from day to day.

Before any capital investment is made in a new system,this system should be tested. There is no easier way thanthis type of simulation to test planned capital investment,and then weigh the investment up against the return oninvestment.

Nomenclature

Cr = Total operatingcost of the end-user group in randsper hour

Csn = The different operational cost of each sub-systemCpn = ProductioncostCmn = Maintenance costfen = External cost factorCcn = Capital cost factorCe = Energycost factorCmd = Maximum demand cost coefficient

Cmp = Maintenance coefficient

Cpx = Energy coefficientCman = Labour cost coefficientTmd = Maximum demand cost (tariff) (rands per kW)Te = Energy cost (tariff) (cents per kW)

P = Tonnage exceeding maximum design tonnage

References

1. )ACKSON, W.B. Electrical load management. Proc. First Group Engineen'ngSymposium, Engineering on the Mines in the 90 'soRandburg. 1986.pp. 7-10.

2. DELPORT,G.). Integrated electricity end-use planning in deep level mines.ph.D. thesis, University of Pretoria, 1994.

3. AOAMIAK,M., Roberts, D., and Ketz, S. A microprocessor based system forimplementation of variable spot pricing of electricity, IEEE Computer

Applications in Power, 1990. pp. 43-48.

4. GRAY, D. Can you afford to only install a SCADA system today?

Instrumentation and Control. 1995. pp. 23-25. .

The Joumal of The South Afncan Institute of Mining and Metallurgy MAY/JUNE 1997 153 ...

The Australasian Institute of Mining and Metallurgy (The AusIMM) is a pre-eminentassociation serving minerals industry professionals in Australia and the Asia Pacificregion.

The AuslMM has a very active publishing agenda - which Includes Monographs,Conference Volumes and Spectrum Series publications

MONOGRAPHS

Over the years, The AuslMM has published a number of world-class monographs.Monographs of note Include:

. The Field GeologISts' Manual

* Australasian Mining and MetallurgyCost Estimation Handbook for the Australian Mining Industry

CONFERENCE VOLUMES

For every conference held by The AusIMM, a volume of the papers presented ISpublished.

SPECTRUM SERIES

ThIS series of publications includes 'How To' handbooks. historical, lolnt publicationswith other kindred bodies etc.

To obtain your copy of The AuslMM's 1997 Publications Catalogue, complete theform below and return It to:

~"The Publications Department, The AuslMM

PO Box 660, Carlton South Vic 3053, Australia.Facsimile: 61 396623662

}C..,.....................................................Name:

................

Position:-

Organisation:-~._~-

Address: ~--- --- ----

Postcode: Country:----

Telephone:--~ Facsimile: Email:

--------

PLEASE SEND ME THE AuslMM 1997 PUBLICATIONS CATALOGUE

~ 154 MAY/JUNE 1997

MONOGRAPH 20Cost EstimationHandbook forthe AustralianMining IndustryEdited by: M Noakes &T Lanz

This book provides simple instructions for calculating capital andoperating costs using graphical or formulation methods. setting out the

steps necessary for the preparation of an estImate to a given or selected levelof accuracy-

It outlines the method by which the cost estimate is developed, shows howthe equipment desIgn criteria are chosen and the equipment sized andprovides guidelines for the costing of the selected plant with prices inAustralian dollars. Proper use would generate a prehJ1Jinary estimate to apre-feasibility study level of accuracy, though much of the data included isto a higher level for unit operations.

Chapters are devoted to estimating capital costs for surface mining,underground hard rock mining, underground coal and soft rock mining,alluvial mining and tailings reclamation. comminution, concentrationprocesses, materials handling, solvent extraction and electrowinning,infrastructure and predevelopment, and operating costs of surface miningequipment, underground hard rock mining. underground coal and soft rockmining, m1l1eral processing, and administration.

Data from 20 different categones of mineral product are also included forrevenue calculations and marketIng.

ISBN: 0949106879 412 pp

Price: Members of The AuslMM .

Student Members

of The AuslMM ...

Non Members .....

1993

$AUD65.00

$AUD40.00$AUD85.00

{I}~

Please send your order to: The Publications Deportment, The AustrolasianInsltute of Mining and Metallurgy (The AusIMMJ, PO Box 660,

Carlton South Vic 3053, Austrolia.Telephone 61 3 9662 3166; Facsimile: 61 3 9662 3662;

Email: publications@auSlmm.com.au.

Please send me- copies of Monograph 20. Enclosed is my cheque for$A-lincluding postage - please refer to table)

or debit my

Bankcard 0 Visa Card 0 Mastercard 0 Diners Club 0 AMEX 0

Card Number: I

Expiry Date: ~ Signature: Date:-

Please send Monograph 20 to:

Name:

Position: Organisation:

Address:

Postcode: Country:

Telephone: Facsimile:

I"""""":&

I"""'" 'H ,.",,"

! DlIanOlY 01 Economy A" IBoob (approxlmalely 4 weeks dehvelY

,

f-I

I

2-3

4 -,

$AUD3X(MI

$AUDX3.IMI

$AUDI2<J.tM)

When olderolg,pleasc nole Ihallhe po.slage and packmg "ale I11l1sl he lIs",1 when oldenng,dl pllhllc"llo", "nd .1""

pre-I,ayouent by cheque in A",tralian dollar> drawn on an

Australian Dank mu.sl hc ,~<civcd with all ",'det.s. c.'edil ,a.dracs""""

",dc.s ",'eaLso

ac<cptahle.

The Journal of The South African Institute of Mining and Metallurgy

Water quality and reclamation management In mining usingbactericides

V.RAslOGI*

Surfactant-based bactericides are able to control acidformation in sulfidic materials such as pyrite andsulfidic ores, thereby helping to abate acid rockdrainage. Bactericides are used in the following twoways: as periodic or continuous spray treatments forthe prevention of acidification of material activelyhandled and moved and in controlled-release formfor reclamation.

Site-specific bactericide treatments are designedfor active sites and reclamation. Results from twoactive coal refuse sites show that acid productioncan be reduced by 88% or more. Results fromreclaimed sites show that bactericide treatment canbe more effective than the addition of alkalinematerials such as limestone and can reduce acidproduction from 3,500 to <30 ppm. The controlled-release-pellets treatment outlasts the life of the

pellets, as shown by a study of a site now 10 yearsold. A single treatment sets up a change in sitemicrobiology to replace acid-producing Thiobacillusferrooxidans bacteria with heterotrophic bacteriaand creates a stable site without the need for furthertreatments. A secondary beneficial effect ofbactericides is on revegetation, and biomass fromtreated areas is almost 10 times that from areas nottreated with bactericide. Depending on theapplication, bactericide treatment can be costeffective upfront and certainly long-term, comparedto the perpetual water treatment that could berequired, even after productive use of the site iscompleted. .

. PresiJ£n, MVTechndogies,Inc., Akron. Ohw,

=- j

Blue Swallow Worldng Group *

The Blue Swallow Natural Heritage Site just outsidethe village of Kaapsehoop in Mpumalanga, is underthreat by a proposed gold mining company called,ironically, the Blue Swallow Exploration andMining CC.

Partners in the Blue Swallow Exploration andMining CC,Paul Czajowski and John Joubert, havesubmitted a proposal to the Department of Mineraland Energy Affairs to reopen an underground mine,their proposed name for which is the Blue SwallowGold Mine. The mine was closed in 1952 after twoyears of operation due to metallurgical difficulties intreating the ore. Their 'pre-feasibility study' includesplans to conduct alluvial or open-cast-mining.

The current area for which they are makingapplication to conduct prospecting and which theyhope to mine, includes six of the most importantBlue Swallow nests, which would be destroyed by

~_..

the proposed activity. For the past ten years therehave been between nine and 12 active breedingpairs in the Natural Heritage Site each year. TheNatural Heritage Site contains the highest breedingdensity of Blue Swallows in South Africa andSwaziland. At time of writing a decision on theproposal had yet to be made. If mining in thissensitive habitat, specifically set aside for theconservation of the Blue Swallow and its habitat, isallowed, the only Blue Swallows likely to survivewould be the Blue Swallow Exploration and MiningCCand the Blue Swallow Gold Mine. The EWT hasmade a submission to the Minister of EnvironmentalAffairs and Tourism, Dr Pallo Jordan, urging him tointervene and prevent this from happening. .

* Courtesy if the Endangered Wild Ljfe Trust.

The ,Journal of The South African Institute of Mining and Metallurgy JMAY/JUNE 1997 155 ~

- !ill]!TIIN Professional DirectoryThis could be yourProfessional DirectoryAdvertisement ,""SooIhAfri<an"""luteofMlning.""""Uu..,

",,"'io""

. 85 mm wide x 45 mm deep. R300 per insertion. R1200 for six insertions paid in advance.Logocan be includedFor your bookings contact the journal Secretary,

The South African Institute of Mining and Metallurgy13th Floor, Cape Towers, 11-13 MacLaren Street, johannesburg, 2001.

P,O. Box 61127, Marshalltown, 2107.Tel: (011) 834-127317, Fax: (011) 838-5923

SPECIALIST CONSULTING SERVICESIN METALLURGICAL ENGINEERING

AND CORROSION

~ ieiale 'D~ PrEngPhD(CantabjDIRECTOR

p.a. Box 11036, Braoklyn, 0011 Tel: 1012) 420-2187; Fax: 1012) 43-2365

. c - Demolition (pty) Ltd

~c;.1.. Specialists in Demolition and Dismantling of Mine Plants and Structures

Extensive Experience in Plants and Sites with RadioactiveContamination

. Demolition of Concrete and Steel Structures and Concrete Foundationsof all Types

. Leaders in Technologically Advanced Blasting Techniques andMechanical Methods for Safe, Rapid, and Cost Effective Results

Tel.: (011) 888-2859 Fax: (011) 888-2757

Paterson & Cooke .Consulting EngineersPO Box 2143, Clareinch 7740, South Africa, Tel: (021) 683 4734 Fax: (021) 683 4168

Slurry Pipeline Systems.Design of hydraulic transportation systems.System evaluation and design audits.Deep mine backfill distribution system design.Technical feasibility and costing studies.Laboratory pipe loop testing capability

CREmECONSULTING ROCK &

MINING ENGINEERS

Or Paul Buddery CEngPrincipal

Specialists in Coal Mining & Soft Rock

Mine, Tunnel & Support Design. Mining Method & Equipment SelectionGeotechnical Investigation Prngrammes. Codes of Practice. Safety &Technical Audits

25 SHOPETH WAYWOODMANSEY TEL & FAX: +44 1482 888118BEVERLEY MOBILE: 0385568896HU17 OTJ EMAIL: pbuddery@creme.karoo.co.uk

SCHOOLManagement-A Challengefor Mining Engineer.12 June 1997, Mintek Auditorium, Randburg

COLLOQUIUM

Technology for Rapid Access30 July 1997, Mintek Auditorium, Randburg

COLLOQUIUM

Experiences with New Mine Legislation30 October1997', MintekAuditorium,Randburg

INTERNA AL CONFERENCEHeavy Minerals '9720-24 October 1997, Holiday Inn Crowne Plaza, DURBAN

~ 156 MAY/JUNE 1997 The Journal of The South African Institute of Mining and Metallurgy

--------- --~--