You are on page 1of 13

USIM PAC 3.

0: New Features for a Global Approach in


Mineral Processing Design
Stéphane Brochot, Marie-Véronique Durance, Jean-Claude Guillaneau and Jacques Villeneuve1

ABSTRACT
Since 1986, BRGM has been developing a powerful process simulation software package,
USIM PAC. These last years have seen significant developments in mineral processing
technologies, particularly in hydrometallurgy, bio-hydrometallurgy and mineral liberation.
The new version of the simulator, USIM PAC 3.0, incorporates these modern-processing
fields. Its structure and tools allow the user to take into account, at the same time, a wide range of
technological, economical and environmental aspects.
The improvement of the existing features and the introduction of new unit operation
mathematical models and tools such as the supervisor, global optimization, workshop model or
chemical reactions contribute to a global approach in mineral processing design.

INTRODUCTION
Since 1986, BRGM has been developing a powerful process simulation software package,
USIM PAC (Broussaud 1988; Durance et al. 1993; 1994; Guillaneau et al. 1997). It is a user-
friendly steady-state simulator that allows mineral processing engineers and scientists to model
plant operations with available experimental data and determine optimal plant configuration that
meets production targets. The simulator can also assist plant designers with sizing unit operations
required to achieve given circuit objectives.
The software package contains functions that manipulate experimental data, calculate
coherent material balances, sizes and settings of unit operations, physical properties of the
processed materials, simulate plant operation and display results in tables and graphs. Widely used
to design and optimize industrial plants, with more than one hundred fifty copies sold in thirty
countries, this software has been continuously improved, through successive versions, to make it
more accurate and easier to use.
These last years have seen significant developments in mineral processing technologies,
particularly in hydrometallurgy, bio-hydrometallurgy (Cézac et al. 1999; Brochot et al. 2000) and
mineral liberation. In addition, it is now necessary to take into account the environmental impact
at each stage of a mining project, including water and power consumption, waste treatment and
disposal (Sandvik et al. 1999; Guillaneau et al. 1999).
This new version of the simulator, USIM PAC 3.0, incorporates these modern-processing
fields. Its structure and tools allow the user to take into account, at the same time, a wide range of
technological, economical and environmental aspects.
After a reminder about the simulation-based approach and the main features of USIM PAC,
new features will be presented. The particular case of the workshop model will be treated. The
interest of these features will be demonstrated through an example.

1
BRGM, Environment & Process Division, Orléans, France

1
THE SIMULATION BASED APPROACH
This approach is using the mathematical description of the unit operation performances to predict
either the design of a new plant or the performances of an existing one under changing conditions
(see Figure 1).

Physical
Physical properties
properties
Input
Input

•• Flowrates
Flowrates MODEL Output
•• Sizes,
MODEL Output (s)
(s)
Sizes,
•• Grades,
Grades,
•• etc.
etc.
Parameters
Parameters

•• Feed
Feed stream
stream
description
description •• Plant
Plant operation
operation
•• Equipment
Equipment •• Plant
Plant capital
capital cost
cost
configuration
configuration

Steady-state
Steady-state
•• Model
Model simulator
simulator •• Plant
Plant operation
operation
parameters
parameters •• Objectives
Objectives
•• Equipment
Equipment
configuration
configuration

Figure 1 Functions of the models and the steady-state simulator

The simulation-based approach may be used for three main purposes. The first one is for the
preliminary design of a new plant, at this stage a balance of the circuit is needed with information
on all the streams (i.e., flowrates, grades, size distributions) and then the design of the main
equipment of the future plant is obtained, including an estimation of the investment costs.
The second one concerns the advanced design of an installation. With more information
coming from lab tests or pilot plant runs, the prediction of the plant operation is more precise and
can be used to study several alternatives of flowsheet or settings, to help at the plant
commissioning stage or be a tool for the plant operators training.
The third classical application is plant optimization (or audit, retrofit and adaptation). In this
case the available data from the plant is used to build a simulator of the plant which can be used,
after validation, to test all the changes which are needed to improve the plant operation.

MAIN FEATURES OF USIM PAC


USIM PAC offers powerful and easy-to-use methods to help engineers reach their objectives. It
requires no special training in computing or modeling.
Basic functions of USIM PAC can be divided into plant modeling, data input, data processing
and different tools to display data and results. High level functions are also available for
configuration and incorporation of user defined functions.

Flowsheet drawing
The first element of the plant model is the flowsheet describing the different unit operations and
material streams. The «Flowsheet Drawing» tool (see Figure 2) performs this description. It is
entirely graphic and icon driven with over 100 unit operation icons available through a toolbox.
Streams and units are referenced by a unique number and can be named by a label, which will
subsequently appear everywhere in the software. Colors and text fonts can be modified to increase
the visual comprehension of the flowsheet.

2
Figure 2 Flowsheet drawing tool

Phase model
In USIM PAC the material flowing into the streams is described in terms of physical phases.
There are three types of phase corresponding to the three material states: solid liquid and gas. A
fourth type named “ore” is defined to identify the raw material from other solids like carbon.
Each phase is described by its flowrate and some description criteria among “particle size”,
“composition” (mineralogical or chemical composition), “particle type”, “floating ability” sub-
populations and “user defined sub-populations”. Two criteria can be crossed to have a finer
description such as mineralogical composition by size class.
A set of predefined phases offers different ways to describe material according to the
available data and the level of complexity imposed by the objectives and the unit operation models
(see Figure 3).
The ability of USIM PAC and its mathematical models to work with a wide range of
predefined or user-defined material phase descriptions is certainly the main feature and strength of
the software. Mechanisms of communication between the phase model and the mathematical
models of unit operations have been largely presented in previous papers (Durance et al. 1993;
Brochot et al. 1995).

3
Ore
Ore
Ore
Ore

Water
Water
Water
Water Carbon
Carbon
Ore
Ore
Solution
Solution
Grinding
and flotation
plant Gold ore
treatment
Preliminary plant
grinding
plant
design

Figure 3 Different phase models

Equipment description
Various mathematical models can be associated to each unit operation drawn on the flowsheet.
Mathematical models calculate the output streams data from the input stream data and model
parameters (see Figure 1). These parameters can be equipment sizes, operating conditions,
physical properties, model adjustment parameters or simply performances. Depending on the
objective of the simulation and the available data, different mathematical models can be used for
the same equipment. In USIM PAC, mathematical models are divided into four levels:

· Level 0 models enable the user to specify directly the performance of the units. For
example, the performance of a classification unit can be modeled by a partition curve for
which the user indicates the bypass, the imperfection and the d50 (the cut-size). Such
models are flowsheeting models that do not take into account any dimensional
parameters. During the simulation, the performance of the unit will be independent of its
dimensions and the flowrate of the ore feeding it.
· Level 1 models take dimensional parameters into account. They require little (sometimes
no) experimental data. A typical example is a ball mill model, which uses only the Bond
Work Index as experimental parameter. If no data is available, it is even possible to
estimate the Work Index. Obviously the precision of such models is limited, but they are
easy to use.
· Models of higher level are more accurate but they require the estimation of some of their
parameters. This estimation can be carried out either on the basis of experimental data
obtained from the continuous operation of the unit (level 2 models) or from such data
supplemented by information obtained from specific tests, generally carried out in the
laboratory (level 3 models).

Over 120 mathematical unit operation models are available covering a wide range from
crushing to refining, from ore dressing to waste management. These include comminution (SAG,
Pebble/Rod/Ball mils, SAM, etc.), classification (Hydrocyclones, Screens, Rake/Spiral classifiers,
etc.), concentration (Conventional/Column floatation, Gravity/Magnetic separation, etc.),
hydrometallurgy (Leaching, Bioleaching, CIP, CIL, Precipitation, Cementation, Solvent
extraction, Electrowinning, etc.), solid/liquid separation (Filtration, Sedimentation, etc.), waste
treatment (Collection, Sorting, Incineration, composting, etc.).

4
Stream data
The input interface for the stream data follows the phase model’s structure. It gives, stream by
stream, the list of phases and solid-liquid pairs (see Figure 4).

Figure 4 Stream data entry

Available data for each phase are the mass flowrate, the volumetric flowrate and the density if
they are known, the stream density if it is required stream by stream and the descriptions such as
size distribution, composition, composition by size classes or floating ability by component.
Available data for each solid-liquid pair are the pulp mass flowrate, the pulp volumetric
flowrate and density if they are known and the solid percent. For a given solid-liquid pair, only
two values among both phase flowrates, pulp flowrate and solid percent are necessary. The other
two are calculated.
Error columns display the confidence level of experimental data used for data reconciliation
by material balance.

Data processing
Classical use of USIM PAC follows three main methodologies as described in previous papers
(Durance et al. 1994): preliminary design, advanced design and plant optimization. These
methodologies require various algorithms that are available in USIM PAC:

· The data reconciliation by statistical material balances (Le Guirriec, Brochot, and
Bergounioux 1995) giving a set of coherent stream data from a set of experimental
incoherent data coming from different kinds of measurement (on-line sensors, different
pilot plant tests, laboratory tests, etc.) with varying levels of confidence. The only rules
used in this algorithm are the material conservation laws.
· The direct simulation is a sequential modular iterative algorithm. It calculates all stream
data from feed data and unit operation parameters. It is possible to perform a direct
simulation only on a selected sub-flowsheet or on a sole unit operation.
· The optimization algorithm searches unit operation parameters giving output streams as
close as possible to objective streams. This algorithm is used for unit sizing, model

5
calibration or physical property calculation, the difference appearing in the set of desired
parameters: dimensions, adjustment parameters or physical properties. The objective
stream can be an entire stream data set or some stream data parameters such as the d80, a
component grade or a recovery. It can also be an output unit operation parameter such as
the power consumption.
· The Objective Driven Simulation (ODS) is a mix between direct simulation and
optimization (Villeneuve et al. 1992). At each simulation iteration, it calculates, for each
unit operation, the parameters reaching a target. This algorithm is generally used to
improve the plant calibration previously done unit by unit using optimization.
· The global optimization use the same algorithm as the unit operation optimization but
applied to the entire flowsheet. It finds parameters of some unit operations that meet a
plant performance objective. For example, it can be used on a floatation circuit to
evaluate the number and volume of cells per bank giving the best compromise between
grade and recovery to maximize profit.
· The capital cost estimation calculates the investment cost of each unit operation using a
cost model and estimates the overall plant construction cost using a hierarchical ratio
system (Mular 1982). The list and level of the ratios is entirely configurable.
· The supervisor can be used either as a sensitivity analysis tool or for visual optimization
(Guillaneau et al. 1995). It calculates user defined plant parameters (global or local
performances, constrained parameters, etc.), named sensors, when some input parameters
(feed data or unit operation parameters), named actuators, are varying. It is then possible
to draw sensor variations as a function of actuators and observe the sensitivity of a plant.
In the case of a multi-criterion optimization, it is easier to choose the best configuration
regarding such a graph than to translate the target into an objective function.

Each algorithm has many options to translate simulation objectives in terms of mathematical
problems. But these algorithms can be easily used thanks to a set of default options chosen for
their ability to respond to the most often encountered situations. This is the reason why
USIM PAC can be used by process engineers as well as by researchers.

Graphics
There are eight distinct graphical representations: size distribution curve, size partition curve,
density distribution curve, density separation curve, split curve, stream bar graph, component bar
graph and supervisor results. These graphs are entirely configurable. Some predefined graphs can
be drawn directly from the flowsheet popup menu. From a stream submenu it is possible to draw
the size distributions of all solid components. From a unit operation submenu it is possible to draw
the size partition curve or the split curve.

Global results
Detailed results are displayed using the same interfaces as for stream and unit data input. They can
also be displayed in an overview sheet, which can be printed or directly exported into a
spreadsheet.
The global result sheet is a summary of results. It displays for each stream -given in lines- a
set of calculated parameters -given in columns (see Figure 5). These parameters are user defined
from a large list of possible parameters. It is then possible to draw a stream bar graph from the
selected column of a global sheet.

6
Figure 5 Global results sheet

Configuration
To cater for the diversity of users (different application fields, countries, etc.) USIM PAC is highly
configurable:

· Icon configuration allows the user to change each icon’s name, position in the icon menu
and attached unit operation models.
· Model parameters configuration allows the user to change, for each model, its name and
the name, the default value, the variation range and the type of its parameters. It is also
possible to create variants of an existing model. It is then possible, using the same
mathematical model, to create a new unit model by changing the name of its parameters
and their default values. It is a mean of constructing a customized equipment database.
· Phases configuration is a tool to create new predefined phases for the phase model.
· Physical units configuration is available everywhere there is a unit. It allows the user to
add new units in the different unit systems to be in accordance with his habits. The
currency configuration constitutes a particular case.
· Graphs and global results configuration gives the default options for these displays.

Development Kit
USIM PAC also allows the operator to add new functions, such as incorporating icons and models
into the standard libraries with the USIM PAC Development Kit. This module offers an
application programming interface for communication between USIM PAC and the user
FORTRAN code.

NEW FEATURES IN USIM PAC 3.0


Compares with the previous version, the file management has been drastically simplified. In
version 3.0, one single file with the extension .UP3 contains all the plant data (flowsheet, phase
model, stream and equipment description). Each algorithm uses data from this file and generates a
new one containing results. The application is now multi-document and can display
simultaneously various plant files, graphics or global results tables.
Globally, all user interfaces have been improved using spreadsheet-like tables that permit
copying and pasting cells.

7
The following sections highlight the most significant new features of USIM PAC version 3.0.

Flowsheet drawing
It is now possible to attach a label to each stream and each unit operation. To improve the
legibility of the interfaces, these labels are displayed everywhere in stream description, equipment
description, global results, stream and unit choice for the curve characteristics of the graphs,
capital cost report.

Phase model
Four new criteria are now available for phase modeling: “particle type” and three “user defined
sub-population.
The criterion “particle type” is used only for solid or ore phases to model mineral liberation. It
is used in conjunction with the “composition” criterion to describe different classes of particles
having different mineralogical composition. Unit operation models identifying individual particles
such as for comminution, size classification, floatation, gravity or magnetic separation use particle
types instead of mineralogical composition. Other unit operation models such as leaching use
mineralogical composition, directly connected to the elemental composition.
There are three “user defined sub-populations” that can be used to follow some material
properties not treated by the unit operation models but required by the user. Taking the example of
a soil contaminated with heavy metals, the treatment will only be concerned with size
classification and density (mineralogical) separation. Heavy metal contents are not necessary for
unit operation models but are absolutely necessary to evaluate the process performance. It is then
possible to add a user defined sub-population criterion in conjunction with the mineralogical
composition, giving the mean heavy metal contents for each stream.
As some unit operation models describe material transformations in terms of chemical
reactions, the phase model description includes all the possible chemical reactions appearing
anywhere in the process. As these reactions are generally given in stoichiometric proportions, it is
necessary to specify the molecular mass of the different components.
If the density of the solid material is easily calculated from the individual component densities
and the composition, it is not so easy to predict the density of a complex liquid phase such as an
electrolytic solution. In this case, it is useful to be able to specify the liquid phase density for each
stream. A new description named “stream phase density” has been added to some predefined
liquid phases for this purpose.
This new version offers the possibility to define solid-liquid pairs that constitute a pulp phase,
which will be displayed in the stream description.

Equipment description
New models have been added in USIM PAC 3.0, particularly in hydrometallurgy, waste treatment
and solid-liquid separation. Some previously available models have been improved for power
consumption calculation or capital cost estimation. In some existing hydrometallurgical models,
chemical transformations are now expressed as chemical reaction equations. For example, this
improvement allows the evaluation of effects associated with leaching such as cyanide
consumption, degradation or co-leaching.

Stream data
In addition to the pulp characteristics, stream data input offers new possibilities, such as size
distribution input in individual, passing or retained form; chemical composition input of solution
in mass grade (%), concentration (g/l), molarity (mol/l), molality (mol/kg of solvent) or molar
fraction (%).

8
Data processing
Even though the supervisor was available in version 2.1+, it is now proposed in the standard
version. The global optimization, which is time consuming, is now possible due to the
improvement in computer speed.
For the optimization algorithms (unit and global optimization, ODS), the main improvement
is the definition of the objective function. In previous USIM PAC versions the only available
function was the sum of the least square differences between all calculated and objective partial
flowrates of all output streams. Now, it is possible to refine the objective function as a weighted
sum of elementary objective functions. An elementary objective function is a comparison between
a calculated stream parameter (d80, component grade, recovery, etc.) and a given value or the
same parameter in an objective stream (see Figure 6).

Figure 6 Objective function definition

THE WORKSHOP MODEL


The workshop model is a particular model available for all icons. It encapsulates a sub-flowsheet
representing a part of the plant.
In the example of the Figure 7, the global plant flowsheet is displayed, showing the different
parts of the flowsheet: grinding circuit, concentrator, extraction circuit, gold production circuit,
water recycling circuit and tailings effluent treatment as a plant building icon.
Icon #1 (Crushing, grinding & gravity) is associated with a sub-flowsheet by choosing the
model Workshop in the Unit of Equipment Description box. The flowsheet is edited by
choosing the item Workshop Editing in the unit operation popup menu. A new window is opened
to display the grinding circuit (see Figure 8).
The phase model associated with this sub-flowsheet will always be exactly the same as the
plant flowsheet. This is required in order to have a simple correspondence between plant and
workshop stream description. The interfaces for stream data input and equipment description are
identical to the plant ones.
The last step in using the workshop is to associate input and output streams of the workshop
icon to the streams of the sub-flowsheet. That is possible using the Unit of Equipment
Description box in which the list of streams linked to the workshop icon is displayed in the first
column. The corresponding sub-flowsheet streams are selected in the second column (see
Figure 9).

9
Figure 7 Global plant flowsheet

Figure 8 Grinding workshop flowsheet

10
Figure 9 Workshop streams association

During a simulation, when the workshop model is called, it copies input stream data into the
corresponding streams of the sub-flowsheet, runs a simulation of the sub-flowsheet and then
copies in the output streams the corresponding stream data calculated by the sub-flowsheet
simulation. In addition to the better legibility of a complex flowsheet, the workshop also allows to
increase the simulation speed. Actually the convergence of the circuit is obtained before passing
onto the following circuit. If the number of recycling loops between the different workshops is
low, the simulation time is significantly reduced.
The workshop model can also be used to model a unit operation that can be described as an
association of several existing models. For example, in the case of cyanide addition before
grinding, leaching appears inside the ball mill. A simple ball mill model and a leaching model
linked in series can represent this unit operation. The ball mill icon is then associated to a
workshop containing a little sub-flowsheet with a ball mill followed by a leaching tank.
It is possible to associate a workshop model with an icon of a sub-flowsheet. There is no
theoretical limit to the number of such levels.

THE GLOBAL APPROACH THROUGH AN EXAMPLE


The example of Figure 7 shows the flowsheet of a gold ore treatment plant. The objective of this
study is the preliminary design of the plant, taking into account the global economical and
environmental aspects.
Attention will be paid to the following points:

· Global water consumption with maximization of water recycling;


· Global cyanide consumption taking into account gold leaching, co-leaching, losses and
recovery by leaching water recycling;
· Global carbon consumption taking into account losses by abrasion and aging.

Water recycling
To evaluate the global water consumption and maximize the water recycling, the simulator
includes a water distribution line. This line consists of a water line regulator (Unit 7 in Figure 7)
that calculates the quantity of fresh water added to the recycled water so that water demand along
the line is met and water losses compensated. The water demand is calculated using other
regulators such as pulp density regulators in the grinding circuit (Units 2, 4, 6 and 8 in Figure 8),
which take the desired quantity of water away from the line.
The remaining water at the end of the line is added to the clarified water coming from the
solid-liquid separation circuit and to the treated effluents coming from tailings. All this water
returns to the line regulator with a different chemical composition than the fresh water. This
system allows the evaluation of the accumulation of some chemical elements due to the water
recycling.

11
Cyanide consumption
Using new leaching models, it is possible to take into account main reactions with cyanide
occurring in the leaching tank. It is also possible to take into account reactions with cyanide
occurring in other unit operations such as CIP.
These models use the chemical reaction-like description now available in USIM PAC 3.0.
This tool describes the material balance during chemical reactions. It is used in conjunction with a
simple equilibrium model organizing chemical reactions into an activating hierarchy. The
reactions are successively activated level by level in this hierarchy and competition between
reactions of the same level is governed by reaction ratios. The reaction limitation is governed
either by the available reagent ratio or by the concentration set-point of the products. The available
reagent ratio is calculated, in the case of the leaching model, using a leaching isotherm.
Using these models, it is possible to take into account not only the main reaction (gold
extraction) but also other reactions in competition such as co-extractions or cyanide degradation.

Carbon consumption
USIM PAC includes mathematical models for CIP taking into account the carbon losses by
abrasion. This fine carbon is lost in the barren pulp and is loaded with gold.
This lost carbon has to be compensated by adding fresh carbon. It is also necessary to replace
regenerated carbon by fresh carbon after a number of cycles. New regulation models simulate
these mechanisms maintaining a constant flowrate and quality of carbon cycle.

CONCLUSIONS
Through this example, the USIM PAC simulation software shows its ability to model and simulate
a wide range of process types and technologies. It offers various tools responding to a multi-
criterion and global approach more and more in demand for a plant design study. It takes into
account different aspects:

· Economical aspect as capital cost, reagent and power consumption, production quality in
terms of valuable mineral grade or undesirable elements level;
· Technical aspect with an evaluation of various configurations and techniques, a complete
and detailed description of all material streams and their behavior during process;
· Environmental aspect as water consumption and recycling, pollutant production or waste
treatment.

USIM PAC could just as well be used by process engineers for plant design or optimization as
researcher for process development or teacher for education.
The last version, USIM PAC 3.0, constitutes an improvement in the integration of different
industries to globalize the approach. It is currently possible to simulate treatment from the mine to
the metallurgical plant. Current works on a global approach in urban waste management (Sandvik
et al. 1999) or metal life cycle (Reuter 1998) using techniques of the steady state process
simulation anticipate the future developments in which the mineral processing will integrate the
industrial chain from raw material production to ultimate waste of products.

ACKNOWLEDGMENTS
This paper is the BRGM contribution n°1690 for which the work was in part financed by a BRGM
research project.

REFERENCES
Brochot, S., M.-V. Durance, S. Foucher, J.-C. Guillaneau, D. Morin, and J. Villeneuve. 2000.
Process simulation to enhance complex flowsheet development: examples in biotechnology.
SME-Control 2000 Conference.

12
Brochot, S., M.-V. Durance, G. Fourniguet, J.-C. Guillaneau, and J. Villeneuve. 1995. Modelling
of the Minerals Diversity: a Challenge for Ore Processing Simulation. Proceedings
EUROSIM’95 Conference. 861-866.
Broussaud, A. 1988. Advanced Computer Methods for Mineral Processing: Their Functions and
Potential impact on Engineering Practices. Proceedings XVIth International Mineral
Processing Congress. 17-44.
Cézac, P., X. M. Truong-Meyer, X. Joulia, S. Brochot, and D. Morin. 1999. A New Modelling
Approach of Bioleaching Process. CDROM of the ECCE2, Second European Congress of
Chemical Engineering.
Durance, M.-V., J.-C. Guillaneau, J. Villeneuve, G. Fourniguet, and S. Brochot. 1993. Computer
Simulation of Mineral and Hydrometallurgical Processes: USIM PAC 2.0. a Single Software
from Design to Optimization. Proceedings International Symposium on Modelling.
Simulation and Control of Hydrometallurgical Processes. 109-121.
Durance, M.-V., J.-C. Guillaneau, J. Villeneuve, S. Brochot, and G. Fourniguet. 1994. USIM PAC
2 for Windows: advanced simulation of mineral processes. Proceedings 5th International
Mineral Processing Symposium. 539-547.
Guillaneau, J.-C., J. Villeneuve, S. Brochot, M.-V. Durance, and G. Fourniguet. 1995. The
Supervisor of Simulation: a step further to meet the Process Engineer Needs. Proceedings
XIXth International Mineral Processing Congress.
Guillaneau, J.-C., J. Villeneuve, M.-V. Durance, S. Brochot, G. Fourniguet, and H. Durand. 1997.
A range of Software for Process Analysis. SME Annual Meeting. Preprint # 97-202.
Guillaneau, J.-C., S. Brochot, M.-V. Durance, J. Villeneuve, G. Fourniguet, H. Védrine, K.
Sandvik, and M. Reuter. 1999. From mineral processing to waste treatment: an open-mind
process simulator. CIM 1999.
Le Guirriec, E., S. Brochot, and M. Bergounioux. 1995. An Augmented Lagrangian Method for
Problems Arising in Mineral Processing. Proceedings 17th IFIP TC7 Conference on System
Modelling and Optimization. Vol. 1, 65-68.
Mular, A.L. 1982. Mining and Mineral Processing Equipment Costs and Preliminary Capital Cost
Estimations. Canadian Institute of Mining and Metallurgy.
Reuter, M.A. 1998. The Simulation of Industrial Ecosystems. Minerals Engineering.
11:10:891-918.
Sandvik, K. L., J. Villeneuve, M.-V. Durance, and H. Védrine. 1999. Development of a Mineral
Processing Program as a tool for optimal decision in Waste Treatment. Proceedings
REWAS’99, Global Symposium on Recycling, Waste Treatment and Clean Technology. Vol.
1, 55-64.
Villeneuve, J., J.-C. Guillaneau, R. Pilotte, and A. Broussaud. 1992. Objective Driven Simulation:
a new Approach to Improving the Efficiency and Usefulness of Steady-state Simulators of
Mineral Processing Plants. SME Annual Meeting. Preprint # 92-168.

13

You might also like