Search Documents

By Donald Leroux, Martin Suichies, Andy Trusiak, Chris Dechert

"In modem concentrators, the operation of flotation machines is often stabilized by simple control loops. Froth level, slurry density and aeration rate are all physical parameters typically controlled through classical PID loops. Few concentrators, however make use of advanced computing technologies for controlling the grade of flotation product streams. Technical challenges such as long process delay and the long and uneven sampling interval of on-stream XRF analyzers often restrain their implementation. Model Predictive Control (MPC) is known to have better performance than PID for controlling processes with long delays (such as flotation circuits). This paper describes the implementation of a multi-input/single-output (MISO) MPC algorithm and an FIR model identification routine using an object-oriented graphical computing environment.The MPC module (controller and identification) was implemented on an industrial control system and applied to many sulfide flotation circuits in the Brunswick Mining concentrator (New Brunswick, Canada). In most cases, the manipulated variable was the collector addition. As long as it is provided with decent process models, the performance of the MPC controller is good on all the flotation circuits on which it was applied. Significant reductions of grade variance are reported and the controller responds well to set-point changes. The user-friendly, graphical nature of the computing environment led to a high degree of acceptance by both the process engineers and the flotation operators.1. INTRODUCTIONOver the past three decades, much instrumentation has been developed for flotation machines. The introduction of instruments such as pulp density gauges and cell level probes was particularly well received by the plants where pneumatic and electronic controllers were available for implementing PID control loops. In a way, these simple loops revolutionized the mineral industry. The operation of flotation circuits could then be stabilized through the use of automatic controllers."

Jan 1, 1999

By Anuj Narang

Large-scale separation or flotation cells, used in the primary extraction in the oil sands industry, are integral parts of the overall process of bitumen extraction. Good regulation of the interface level between the bitumen froth and the middlings in these cells can result in a significant improvement in bitumen recovery and throughput, and heavily influence process economics. Automatic control of such industrial processes presents challenges such as multi-phase flow dynamics, variable process gain and time delays, poor quality of measurements and significant disturbances. This paper details a case study application of development of multiple models of the dynamics of the interface level and the design of model predictive controllers (MPC) for a separation cell process. A Model predictive controller (MPC) using linear models is designed and implemented in real time on an industrial separation cell. The industrial implementation results show that the MPC scheme provides significant benefits over the current operations which use a PID controller. The benefits include significant reduction in the variance of the interface level resulting in higher bitumen recovery and smoother downstream operation plus energy savings due to reduced underflow pump speed.

Sep 1, 2012

"This model specification includes sample text shown in normal font with commentary in italics. The commentary is intended to highlight project-specific items the specification writer should consider during development of the specification. In some sections, typical values are presented to indicate common practice, and these values may be changed to reflect projectspecific requirements. The commentary should be removed before using the specification for a project.1. SCOPEThe work consists of designing, furnishing, installing, post-tensioning and testing helical anchors used to support tension loads, and any ancillary materials (e.g., load transfer devices, etc.) according to the project plans provided and these specifications. Unless otherwise noted, the installing contractor shall provide all labor, tools, equipment, and materials necessary to accomplish the work. The parties referred to in this specification are as follows:• The Owner is the property owner.• The Owner’s representative is the individual or firm hired by the Owner to execute project tasks required for the successful design and construction of the overall project. Tasks could include construction management, design, oversight, etc.• The anchor designer is the individual or firm generally hired by the installing contractor to design the helical anchors.• Installing contractor installs and tests (if necessary) the helical anchors, and possibly other tasks associated with the project.It is important to clearly define the roles and responsibilities of every party on each project to promote good communication and reduce potential for contractual misunderstandings. Different types of contractual vehicles (design-build, value engineering, etc.) require different roles and responsibilities for the various parties, which should be considered in specification development.The Owner will provide suitable access to the construction site for the installing contractor’s personnel and equipment. Unless specifically noted otherwise in the contract documents, the Owner will remove and replace any structures, utilities, pavements, landscaping or other surficial improvements in the work area as necessary to facilitate the work. The Owner will be responsible for overall construction oversight to minimize potential for developing unsafe conditions. The work may include post-construction monitoring of anchor performance if required by the Owner or the anchor designer to confirm performance of the anchors subsequent to their installation. The work may also include installation of additional anchors if the initial installation does not meet specified capacity or performance requirements. Responsibility for providing subsurface investigation shall be addressed in the bid documents. The Owner is responsible for providing to the installing contractor all available subsurface soils, rock and groundwater information and geotechnical testing data."

Jan 1, 2016

"This model specification includes sample text shown in normal font with commentary in italics. The commentary is intended to highlight project-specific items the specification writer should consider during development of the specification. In some sections, typical values are presented to indicate common practice, and these values may be changed to reflect project-specific requirements. The commentary should be removed before using the specification for a project. 1. SCOPE The work consists of designing, furnishing, installing, and testing helical piles used to support foundation loads (compression, tension and lateral), and any ancillary materials (e.g., load transfer devices, etc.) according to the project plans provided and these specifications. Unless otherwise noted, the installing contractor shall provide all labor, tools, equipment and materials necessary to install the required piles. The parties referred to in this specification are as follows: • The Owner is the property owner. • The Owner’s representative is the individual or firm hired by the Owner to design and oversee the overall project. • The pile designer is the individual or firm generally hired by the installing contractor to design the helical piles. • Installing contractor installs and tests (if necessary) the helical piles, and possibly tasks associated with the project. It is important to clearly define the roles and responsibilities of every party on each project to promote good communication and reduce potential for contractual misunderstandings. Different types of contractual vehicles (design-build, value engineering, etc.) require different roles and responsibilities for the various parties, which should be considered in specification development. The Owner will provide suitable access to the construction site for the installing contractor’s personnel and equipment. Unless specifically noted otherwise in the contract documents, the Owner will remove and replace any structures, utilities, pavements, landscaping or other surficial improvements in the work area as necessary to facilitate the work. The Owner will be responsible for overall construction oversight to minimize potential for developing unsafe conditions. The work may include post-construction monitoring of pile performance if required by the Owner or the pile designer to confirm performance of the piles subsequent to their installation. The work may also include installation of additional piles if the initial installation does not meet specified capacity or performance requirements. Responsibility for providing subsurface investigation shall be addressed in the bid documents. The Owner is responsible for providing to the installing contractor all available subsurface soils, rock and groundwater information and geotechnical testing data."

Jan 1, 2015

By D. K. Panda

The behavior of single drops with gas evolution at the interface is of interest because this is one of the important reactions that occur in oxygen steelmaking vessels, where a large amount of iron-carbon drops are dispersed in an oxidizing slag. The carbon dissolved in these drops, reacts with the slag to form CO, which is evolved as a gaseous phase. A room- temperature model was used to study the nature of this type of reaction including the effect of various factors such as viscosity and the presence of surface active impurities. Experiments were done for reacting drops over a wide size range to observe various regimes of behavior. Large single drops were levitated in a counter current flow to produce a sufficiently long residence time to allow the study of the various stages in the reaction. The emphasis of this work was to obtain photographic evidence of a high quality to describe all possible modes of behavior, and so a lot of effort was given to developing and perfecting experimental techniques.

Jan 1, 1994

By S. Kang, D. G. C. Robertson

Slag cleaning furnaces require extensive wall cooling because they involve the use of a superheated slag bath. This is in direct contrast to smelting furnaces, where the wall is usually (but not always) protected by "banks" of solid charge. In order to understand the heat transfer and flow in slag cleaning furnaces a number of model studies have been carried out using molten wax to simulate the slag. Wax is a good model fluid because, like slag, it has a high Prandtl number. Wax models cannot simulate the behavior close to the electrodes (DC arc or AC submerged), but they can and do give a good simulation of the flow and heat transfer at the wall. An understanding of these phenomena is of great importance for the design of cooling panels in the walls of the furnace. Since superheated molten slag will eventually wear away any refractory bricks over time, the wall cooling must be designed to form a frozen layer of slag, referred to as a "freeze lining". The heat flux density (Q/A) required to do this is given by Q/A = h* ?T, where h is the heat transfer coefficient in the slag phase at the wall, and Li T is the superheat (bulk slag temperature - liquidus temperature). Both Q/A and h vary greatly with position and both are usually a maximum at the slag line. In the experimental work, molten wax baths were heated by block heaters to provide similar energy densities to those encountered in commercial slag cleaning furnaces. This procedure was considered to be valid since most of the electrical energy into a furnace bath is dissipated as heat. Gas stirring was used to simulate the gas evolved at the electrode in real furnaces - the quantity of this gas could be estimated from a knowledge of electrode consumption. The similarity criterion was equal gas flow per unit area of surface, in model and prototype. Heat transfer coefficients were measured in the models as a function of power density, gas stirring, superheat, etc., and could be related to the full-scale operations by the use of dimensional analysis.

Jan 1, 1999

By R. C. Hirschfeld, H. H. Einstein, R. W. Bruhn, R. A. Nelson

The objective of the model studies described in this paper is to determine the effect of planar discontinuities on the strength and deformability of a rock mass. A model material was used because it satisfied concurrently three requirements that available natural-rock samples did not: (1) low strength, which was desirable for minimizing the loads that had to be applied in the laboratory; (2) ease of producing test specimens with geometrically regular joint patterns; and (3) uniformity of the test specimens. The model tests were performed on specimens made from a single modeling material and tested under triaxial stress conditions with [u,=u,]. Results are presented in this paper for tests on specimens that had (1) a single joint set inclined at several angles to the major principal direction, including 0º and 90°, and (2) two joint sets that were mutually perpendicular, one set perpendicular to and one set parallel to the major principal direction. Several different joint spacings were used in each series. Future tests will include a series with two sets of joints that make oblique angles with each other and with the major principal direction. The results of model tests and tests on cores of jointed rock by several investigators (e.g., Müller and Pacher,l Moore,2 Krsmanovic and Milic,3 Hayashi,4 Rosenblad,5 Jaeger,6 and Lane and Heck 7, have contributed significantly to the design of these experiments, which are believed to be the first in which a comprehensive study has been made of the combined effects under triaxial loading of joint spacing and joint orientation (Fig. 1).

Jan 1, 1970

By Szwedzicki T

Uniaxial testing was performed on model pillars to determine the effect of width to height ratio on pillar and surrounding rock mass stability. Samples were modelled in a shape similar to that of dumbbell weights, with the centre representing

Jan 1, 1994

By K. Krishnapisharody

The phenomenon of slag droplet formation under the action of gas-stirring in steel ladles has been analyzed with experiments conducted in a thin-slice room-temperature model using water and mineral oils as the working fluids. Visual and video investigations have been performed on the interface behaviour along with measurements of liquid velocities in both phases, using the Particle Image Velocimetry (PIV) technique. The measured velocity data and observations on the interface break-up establish that the droplet formation results primarily from an instability mechanism at the interface.

Jan 1, 2008

By G. B. Misra

While investigating on the aerodynamic resistance of airways supported with peripheral timber sets, at regular intervals, the following theoretical equations were developed by the author to estimate the resistance coefficient of such airways: [ ] for S < 1, where f is Darcy-Weisbach resistance coefficient of the airway, C is modified drag coefficient of the supporting member, D is equivalent diameter of the bare airway, 8 is ratio of the approach velocity over the sets to the average velocity of the bare airway, A is cross-sectional area of the bare airway, a is projected frontal area of the sets, A., is cross-sectional area of the air stream at the vena contracta inside the set, S is spacing of the sets, f, is resistance coefficient of the bare airway, l is length of aerodynamic influence of sets, p is perimeter of the bare airway, p, is setted portion of the perimeter of the bare airway, pe is unsetted portion of the perimeter of the bare airway, and P shielding factor. The equations were verified experimentally in a model rectangular airway supported with one- (bars), three-, and four-piece sets of square-section timber of three different sizes and were found to hold true. The work has been further extended to one-, three-, and four-piece sets of round timber of 2.6, 3.2, and 3.8 cm diam with the same experimental set up. Tests have been carried out for spacings of 25, 50, 75, 100, 150, and 200 cm over a regime of flow defined by the Reynolds number (with respect to the equivalent diameter of the bare duct) ranging from about 1.5 X 106 to 5 X 106 using the same experimental techniques. The values of f are calculated in the manner indicated in [Ref. 1]. Unlike with square-section timber, the resistance coefficient f of the airway setted with round timber shows a distinct variation with the Reynolds number of flow. This conforms to observations made by Sales and Hinsley.2 In order to have a comparable value of f for all types of sets with all sizes of timber, it was necessary to select the value of f at a fixed Reynolds number of flow. Since f is chiefly a function of the drag coefficient of the sets, the appropriate Reynolds number RE is that with respect to the diameter of timber in the set. Considering the diameters of timber used and the regime of flow over which measurements were made, f was chosen at a value of RE = 20,000 in all cases. The f vs. S curves are maximal in nature and in conformity with theory, the f vs. 1/S curves are straight lines up to a value of S = 1 beyond which they show a distinct flexure. The observed values of 1, the length of aerodynamic influence of sets, agree with the relation 1 = 42 e, developed for square-section timber sets, thus suggesting that the shape of timber has little influence on the length of aerodynamic influence. The value of the modified drag coefficient CD for round timber was calculated in the same way as for square timber in Ref. 1, taking the contraction factor Z = 1.5 for round-edged constrictions. CD has an average value of 0.96 with a standard deviation of 6.08% as compared to the free stream drag coefficient of 1.2 at RE = 20,000 for long cylindrical obstructions The shielding factor [ ] is plotted against S/1 in [Fig. 1]. The curves are more or less independent of the size of timber, but are different for the different types of sets, possibly due to their different degree of symmetry. Values of f calculated by the author&apos;s [Eqs. 1 and 2], using experimental values of CD&apos; and [ ] and taking I = 42 e, are plotted in [Fig. 2] against experimentally measured values of f for different types of sets with different sizes of round timber. The values agree closely with a standard deviation of only 5%, thus establishing the veracity of the theoretical equations developed by the author for round timber as well. A comparison was made between the Xenofontowa4 equations (the only other reasonable relations available for the estimation of the resistance coefficient of supported airways) and the author&apos;s [Eqs. 1 and 2] by comparing in [Fig. 3] the values of the resistance coefficient f computed by the Xenofontowa relations with those experimentally measured by the author. In order to make

Jan 1, 1975

By G. B. Misra

Though resistance to flow offered by mine airways supported with timber sets has been an object of study over the last half century, no accurate relation has yet been established from which prediction of the resistance coefficient of a projected timbered airway can be made. Zenofontowa et al. developed an empirical relation, but this relation has not been found to satisfactorily correlate experimental data obtained by Sales and Hinsley. This led the author to develop a set of simple theoretical relations between the resistance coefficient of timbered airways and the various parameters affecting it, such as the type and spacing of sets as well as the shape and size of members forming the set. Experiments were designed with the help of a laboratory model airway to verify the validity of the relations developed. With the experiments conducted so far with 4, 3 and 1-piece sets of 3.4 x 3.4, 2..5 x 2.5, and 4.5 X 4.5-cm square-section timbers in a 45.8 x 38-cm model airway, the theoretical relations have been shown to hold.

Jan 1, 1972

By A. W. Khair

Surface subsidence is a phenomenon mainly associated with longwall mining, although it can also occur if the coal pillar left to protect the surface deteriorates with time and results in the collapse of the room and pillar system. A method has been developed that utilizes physical mine models and laser holographic interferometry. This method is capable of measuring displacements as small as 5 micro- inches by means of generating displacement contours on a surface, even that of granular materials. Analysis of subsidence in room and pillar mining using models of various extraction ratios ad overburden depths along with two different types of overburden model material (sand and gelatin) has been made. The sand represented the extreme brittle conditions of geologic material while the gelatin represented visco-elastic behavior of geologic material. This paper deals with the analysis of data which has been gathered up to date, and mainly associated with the use of sand as an overburden model material. The analysis included the effect of variation of geometric parameters on surface subsidence. Among these parameters are: variation in seam depth and thickness, surface and seam inclination, and extraction ratio.

Jan 1, 1984

By Zhansen P

This paper describes the test re- sults which give a complete proccess of failure of concrete lining embedded in sand medium under load, including the order of appearance of crack along the lining up to failure. nor each crack appearance corresponding calculation being made to assess the redistribution of moment and thrust force in critical section including crown section, section making 45¦ with the vertical etc. De- flection at the above stated critical sections are also measured at each stage of loading. The pressure along the top hori- zontal surface of embedded lining and the vertical sides are measured by pres -sure gauges embedded in sand medium. fhe test results show a complete figure of rupture characteristics of concrete lining embedded in soft ground which approximates alluvial deposite often occur in metro construction at cities near the river side. These test results have been taken as a reference in assessing the action behavior of concrete lining in certain cities in China.

Jan 1, 1987

By R. V. Ramani, V. T. Burgos, J. A. McClay

Model testing is a two part operation; first, to assure the valid¬ity of the mathematical statements and second, to evaluate the accuracy of the computer model for predictive purposes. Verification of both the conceptual and computational aspects can be achieved through the use of input data for which output results are available for comparison. In this dissertation, two test problems were selected for model testing from the Economic Evaluation Reports of the U.S. Bureau of Mines, Process Evaluation Group. Verification Data Set I Input data for this test was selected from an economic analysis of a hypothetical 1,056,000-ton-per-year mine in a 48-inch bituminous coal seam (Process Evaluation Group-Report No. 72-27, 1972). The mine was scheduled to operate five days per week, three shifts per day, 220 days per year with an estimated life of 20 years. Coal reserves covered an area of 6.4 square miles with an expected recovery rate of 70%. Production was achieved through 4 continuous miner sections with an average production rate of 400 tons per shift. The cost summary in the report included a breakdown of equipment and labor requirements. A depreciation schedule listed the expected capital expense per year of operation, assuming capital replacement with identical units and no change in the unit cost. Supply and oper¬ating costs were listed in both gross annual figures and cost per ton of production. The operating and investment costs were projected over the 20 year mine life and a discounted cash flow analysis performed to determine

Jan 1, 1975

By F. Vasilikou

Commercial deployment of Carbon Capture Utilization and Storage (CCUS) requires field testing of a scale that can stress the geologic reservoirs. One such reservoir of interest consists of unminable coal seams that exhibit favorable characteristics and depositional environments and lower pressure and temperature than other, deeper, reservoirs. Such conditions can reduce compression costs while utilizing the action of adsorption that offers a more effective carbon dioxide (CO2) bonding than free storage or solution. To ensure the success of such tests, a number of parameters should be accurately determined, such as the seam geometry and stratigraphy, coal porosity and permeability parameters as well as optimum injection conditions. In essence the CO2 injection model should be calibrated for the reservoir characteristics using all available data. History matching calibrations are based on gas and water production from existing wells prior to injection. This paper will present reservoir models that were developed for a pilot--scale 907 tonnes (1000 tons) CO2 injection test that was performed in 2009 through one legacy coalbed methane production well in Russell County, Virginia, USA. The model incorporates a number of individual coal seams about 0.3 m (1ft) in thickness located at depths ranging from 300 to 700 m (1,000 to 2,200 feet). Model calibration was performed through history matching to prior production data, and subsequently the model was utilized to develop different injection scenarios. The developed model was used after injection to calculate CO2 plume distribution patterns that were monitored at the injection test. The paper will present model data and assumptions with a special emphasis on the effect of hydraulic fractures and the skin factor to the coalbed methane (CBM) production model.

Aug 1, 2013

By B. Bergmark, O. Guyot, A. Broussaud, C. Bouché

The Boliden Aitik concentrator in Sweden processes 2,000 t/h (2,200 stph) of low-grade copper ore. The plant consists of five grinding lines that feed four bulk-flotation banks with a scavenger-concentrate regrind mill und a column cleaner circuit with a regrind mill. The authors installed a model-based expert optimizing control system to maximize copper recovery in the bulk-flotation circuit while maintaining an appropriate copper grude in the cleaner feed. The optimizing control system uses on-line assays and calculates and implements set points for cell pulp levels, air rates and reagent- addition rates. It combines the power of a dynamic population-balance model that estimates mineral liberation in each stream, an optimizer that computes appropriate set points and a rule-based module that implements the strategy. The benefits have been evaluated over a period of three months. The model-based expert optimizing control system generates a measurable increase in copper recovery, around 0.5%, and an optimized bulk-concentrate grade. The need to maintain the optimizing control system and other practical aspects are discussed.

Jan 1, 2000

The Boliden Aitik concentrator in Sweden processes 2,000 tph of low grade copper ore. The plant consists of five grinding lines feeding four bulk flotation banks with a scavenger concentrate regrind mill, and a column cleaner circuit with a regrind mill. The authors installed a model-based expert, optimizing control system to maximize copper recovery in the bulk flotation circuit while maintaining an appropriate copper grade in the cleaner feed. The optimizing control system uses on-line assays and calculates and implements set points for cell pulp levels, air rates and reagent addition rates. It combines the power of a dynamic population balance model which estimates mineral liberation in each stream, an optimizer which computes appropriate set points and a rule-based module which implements the strategy. The benefits have been evaluated over three months. The model-based expert, optimizing control system generates a measurable increase in copper recovery, around 0.5 per cent and an optimized bulk concentrate grade. The need to maintain the optimizing control system and other practical aspects are discussed.

Jan 1, 1998

By André Desbiens, Jocelyn Bouchard, René del Villar

"The metallurgical performance of the column flotation process is determined by the concentrate grade and recovery. Although the former can be continuously measured using an on-stream analyzer, the latter must be estimated by a material balance calculation based on seldom verified steady-state assumptions. Consequently, the automatic control and optimization of flotation columns should be performed using secondary variables having a strong influence on the metallurgical performance, such as froth depth, gas hold-up or bubble surface area flux and bias. The aim of this paper is to present a model-based automatic control strategy using froth depth and bias measurements obtained from virtual sensors developed at Laval University. Preliminary pilot plant results on two-phase (water– air) and three-phase system (water – minerals – air) will be presented and eventual industrial implementation of such a strategy as well as the possibility of achieving effective real-time optimization from bias and froth depth measurements will also be discussed.INTRODUCTIONFlotation columns have been widely studied and used in mineral processing for about twenty years. Although a great deal of work has been accomplished to understand the influence of different variables such as froth depth (H), bubble surface area flux (Sb) and bias (Jb) on the metallurgical performance, the measurement of some of these variables is still difficult to be performed – except at the laboratory level – because in most cases, no commercial device yet exists.As a result of this, it is almost impossible to achieve neither efficient control nor optimization of the process at the industrial level, even in open-loop operation. This paper aims at presenting a simple control strategy of froth depth and bias. Description of the sensors used to achieved measurements, preliminary pilot plant results for froth depth control, eventual industrial applications and possibilities for process optimization are also treated."

Jan 1, 2003

By S. Banisi, M. Pourkani, A. R. Laplante, A. Partoazar

"Canadian experience in industrial modeling of grinding, classification, and flotation kinetics was applied at the Sarcheshmeh copper mine, which processes 40,000 t/d of ore at 0.9 % Cu, in an effort to improve capacity and metallurgical performance. Mcivor's functional performance analysis was applied to ball mill cyclone circuits to maximize circuit capacity. Plitt's cyclone model was applied to predict how apex wear would affect classification performance, as part of a predictive maintenance program whereby the optimum wear period is determined for cast iron apexes. The flotation flowsheet was improved using Agar's optimization criteria. The impact of these technologies has been nothing short of spectacular, with an improved capacity of 1 S to 20""/o, and increased recoveries.1-IntroductionThe Sarcheshmeh copper ore body, which may rank as third or fourth largest in the world, contains 1 billion tonnes averaging 0.90 % copper and 0.03 % molybdenum. It is located southeast of Iran and has been processing 40,000 t/d of ore since 1982. The mine produces 100,000 tonnes of copper and 2,200 tonnes of molybdenum concentrate per year. After three stages of crushing, the ore with a Fso of 12. 7 mm is fed to 8 parallel ball mills in closed circuit with cyclones to produce a product 70""/o finer than 7 S microns. The concentrate of the first stage of flotation is reground and the tailings constitute the main component of the plant final tailings. A concentrate with an average grade of 31 % Cu is obtained after cleaning and re-cleaning stages. Total recovery of the plant varies between 83 and 87 % depending on the operating conditions and ore type.To improve the overall performance of the plant, it was first decided to evaluate the performance of grinding, classification and flotation. Measures were then taken to improve performance using the following criteria: (i) The improvements should be of the types that have resulted in significant improvements in other plants (e.g. Shannon et al, 1993); and (ii) The proposed improvements must be achievable with the manual operation presently used at Sarcheshmeh. The outcome of this study could also help in making decisions on the implementation of automatic control systems in the Sarcheshmeh copper mine."

Jan 1, 1998

By Sushanta Majumder

The physico-chemical characteristics of raw materials used for the production of iron ore pellets have a strong influence on the productivity of pelletization operations and on the quality of pellets. Since iron ores from different sources are used for making iron ore pellets and since the quality of the ores varies with time, there is a need to optimize and control pellet plant operations on a regular basis to achieve optimum productivity while maintaining pellet quality within the desirable specifications. In order to achieve this objective, a model-based on-line optimization system was implemented on two industrial scale induration furnaces. The optimization system consists of a rigorous mathematical model of the induration process, a model for the permeability of the bed of pellets, a statistical model for the quality of iron ore pellets and an optimization module based on sequential quadratic programming. The rigorous model is tested on-line using the wind-box temperature available real-time. The process was optimized to enhance induration furnace productivity for the given constraints on pellet quality parameters such as the cold compressive strength (CCS) and Tumble Index (TI), and other equipment design parameters. Process variables such as induration machine speed, cold air flow rate and hood temperature profile were used as the optimization variables. The on-line optimization system was successfully implemented in two industrial induration furnaces and lead to enhanced productivity. Details of the on-line optimization system and results of process optimization will be presented.

Sep 1, 2012