Search Documents

By E. T. Turkdogan, P. Grieveson, J. F. Beisler

Jan 1, 1963

By K. E. Merklin, F. D. DeVaney

Students of the modern blast furnace seem unanimously agreed that they are observing a major revolution in practice. Rather than changing construction and operation of the furnaces, most of the great advances now under way deal with the raw materials feed. Prepared burdens have gained increasing significance during recent years. Sizing, concentration, and agglomeration have all become bywords to greater production and reduced operating costs. The remarkable results with test furnaces in which self-fluxing sinter makes up part or all of the ore charge foretell even greater improvements. Other investigators have reported many of these advances, and it is enough to predict here that more and more operators will turn to the use of self-fluxing sinter. This change in the operator's requirements comes at a time when one major section of the iron ore regions, the Mesabi, is finding it difficult to maintain, let alone improve, the tonnage and grade of its shipping product. To meet the new demands with respect to quality and structure, the iron ore producers of this area turned first to screening, crushing, and concentrating intermediate ores. More recently, declining higher-grade reserves have led to multimillion-dollar investments in plants to treat the low-grade magnetic taconites.1,2 To yield a suitable concentrate, the Minnesota magnetic taconites must be ground to about 80 pct through 325 mesh, too fine to be charged directly to the blast furnace or to make acceptable feed to the sinter strand. The pelletizing process was developed to meet the structure requirements laid down by the blast furnace operators. Taconite producers are succeeding in their efforts just when the attention of the blast furnace man is being drawn to the great advantages of self-fluxing sinter.3 The taconite operators are ready with an alternative. Why not produce sclf-flus ing pellets? At the Hibbing laboratory of Pickands Mather & Co., where much of the early work was done that led to Erie Mining's 7.5-million ton taconite plant at Hoyt Lakes, Minn., preliminary investigations have been made. Present Taconite Operations: Details of this work will be clarified by a brief review of agglomeration practice in commercial taconite plants. After the taconite has been ground to about 80 pct through 325 mesh to liberate the magnetite grains, the ore is concentrated on magnetic separators, thickened, and filtered to about 10 pct moisture. This structure and moisture fit in closely with the conditions required to produce good balls. It has been found with taconite concentrates that at least 60 pct of the feed to the balling drum must be —325 mesh to obtain good ball compaction within reasonable time. The filter cake moisture usually permits a slight water addition at the balling drum, and further balling control is obtained by adding about 12 Ib of bentonite and 1 1/2 lb of soda ash per dry short ton of concentrate. These additives, and occasionally small quantities of fine coal, are thoroughly mixed with the filter cake ahead of balling. The swelling property of bentonite improves the green and dry strength of the balls so they can be handled without breakage until they are fired. Soda ash is added for pH control to insure maximum performance of the bentonite, and coal is added for more heat in the firing process when required. The balls, normally 3/8 to l-in. diam, are brought to a temperature of 2350" to 2500°F in the indurating furnace. Heat for this process comes from three sources: 1) fuel burned in the combustion chamber, 2) oxidation of the magnetite concentrates to hematite, and 3) coal added to the balls. Pellet strength is developed by grain growth and bridging and, if enough slag-forming constituents are present, by slag bonding. Conversion to hematite is usually close to 100 pct in the fired pellets. Batch Test Procedures: Initial studies of self-fluxing pellets were made by producing balls on a batch basis under carefully controlled laboratory conditions. The l-in. balls, uniform in size, were fired in an oxidizing atmosphere in a standard laboratory muffle furnace. The firing procedure consisted of

Jan 1, 1961

By Yoginder P. Chugh

This study has appraised available deeper soil loosening equipment in the 1-4 ft. depth range, in the U.S. and abroad which has potential for Illinois prime farm-land reclamation. The appraisal was based on written and/or oral contacts with equipment manufacturers, reclamation experts, and visiting with selected equipment manufacturers. The results indicate that several manufacturers in the U.S., West Germany, and England market static and vibratory deeper soil loosening equipment which may have potential for application in Illinois. The pertinent information concerning each equipment with potential is presented in this report.

Jan 1, 1983

By Kenneth Wardell

Criteria have been fully developed for determining the optimum size of oil and gas well blocks. The criteria lead to the formation of well barrier pillars which are, of themselves, inherently stable and take account of variable mine roof and floor conditions. The criteria ensure, through considerations of ground movement, that strains transmitted to well casings, are within tolerable limits of casing specifications. The criteria cater for well barrier pillars in multi-seam mining situations and where previous old mining has taken place. The criteria maintain flexibility in being adaptable to local mine design situations. The criteria are applicable to proposed extraction thicknesses up to 15 feet at depths up to 2,400 feet. The criteria are related to producing wells and to abandoned wells. In the latter connection, various states of well abandonment are recognised.

Jan 1, 1980

By G. S. Scott

In the course of the Bureau of Mines investigation relative to the causes, behavior, and control f anthracite-mine fires, attention was directed toward the correlation between the temperature of oxidation and the gaseous oxidation products in the range of temperatures between 100° and 350°C4/ These tests, made on a high-grade anthracite to determine the amount of oxygen consumed and the amounts of carbon dioxide, carbon monoxide, hydrogen and methane evolved when air was passed through samples of the anthracite at various elevated temperatures, showed that the gases evolved and their relative proportions vary with temperature, constitute a means for determining the probable temperature of the anthracite, and thus permit detection of incipient heating before a dangerous fire hazard exists. As the relationships mentioned are affected somewhat by the amount of previous oxidation of the coal, it was deemed desirable to know to what extent they might vary owing to this cause. It was also considered desirable to have a means for estimating the extent of oxidation from the gas analysis, as this would permit a more accurate estimate of probable temperature. Finally, it seemed desirable to know how much the composition of coal might vary over an extended period of oxidation, since this was previously suggested5/ as a possible means for estimating the probable maximum oxidation temperature to which a sample of coal had been subjected.

Jan 1, 1940

By B Crossley

Self-cleaning fluidised bed heat exchangers have demonstrated unparalleled performance internationally for the implementation of indirect heat transfer in severely fouling fluid environments. Recent advances in equipment design have increased the scope of application to include indirect heat transfer in slurry environments, including fouling, scaling and viscous slurries. The ingenious configuration of recirculated cleaning particles through the tubes of vertical shell and tube heat exchangers provides the ability to solve heat transfer problems that arise as a consequence of tube-side fouling during indirect heat transfer to process fluids or slurries. In most cases, the cleaning particles consist of chopped metal wire, however other process-specific materials may be used.

Jan 1, 2004

By Eugen Gluck, Gus Van Weert, Peter Kondos

"Microwaves have the capability to heat up sulphides, especially clusters of particles of approx. 1 mm. Some sulphides absorb microwaves more readily than others. Molybdenite is especially effective in this regard, possibly due to its flaky nature and/or electrical conductivity. Even where it occurs in low concentrations in a deposit, individual rocks may be ore grade due to the high monetary value of MoS2. Such rocks often result from crushing mined veinlets or stockwork, or are contained in the cutoff grade fraction of ROM. Machine segregation of rocks high in MoS2 by infrared sensors after microwaving could therefore advantageously be used to upgrade lower grade ore between Mine and Mill. This would reduce the size of the latter, and thus its CAPEX and OPEX. Such upgrading could also yield shipping ore for a distant Mill, or extend ore reserves.Barrick Gold Corp. has been supporting the development of MW/IR sulphide sorting for the past three years. The results of MW/IR sorting work on molybdenite and other (pyritic) ore samples from a variety of sources are being presented here.INTRODUCTIONThis paper looks at mechanized, i.e. machine ore sorting technologies currently available and under development, and what conditions would be needed for the potential application of such ore sorting to sulphide mining and milling. Current machine sorting involves no-contact sensing technology for rocks, and signal processing after sensing to activate airblast segregation of certain rocks, normally the smallest mass to keep compressed air costs down.The forerunner of modern sorting is probably magnetic separation, which utilizes an inherent property of certain ores to divert it into mag and non-mags fractions. Radiometric ore sorters, which have been applied successfully for decades in the pre-concentration of uranium ores (Wotruba and Riedel, 2006) initiated modern sorting. In the 90’s, sensor-based machine sorting became a viable means of increasing the overall economic recovery of a resource. This technology became driven by the availability of low cost computing equipment and the concurrent development of algorithms, and by the availability of extremely fast response sensors and air actuators, allowing the sorting equipment to run faster, dropping the cost per tonne. It has been in industrial minerals processing that rock sorting has been most widely adopted, not really surprising. Liberation occurs on a different scale than in the base metal industry. Most industrial minerals are mined from massive deposits, requiring (in the past) no upgrading. Many of these operations now run into product contamination problems, which ore sorting can solve. One illustration: in Finland, a quarry is separating limestone from dolomite at 500 tonnes / hour, with individual rocks weighing up to 15 kg. Many of these industrial minerals operations have stockpiles of earlier discarded below-quality rocks, now being reprocessed with the help of ore sorting. The economics of this are very favourable, going from zero or negative value (stockpiles can not be left when closing the mine) to full product value for at least a portion of the stockpile. This suggests that in the future the economics of overburden sorting could become attractive under the right circumstances and such treatment could well become an integral part of open pit mine development."

Jan 1, 2009

By B. Macdonald

In 1991, roofbolts were introduced as a primary support for inseam retreat longwall entries at CBDC's Phalen Colliery. Increasing mine depth (+550 m) necessitated replacing conventional inseam steel sets in these entries because they were being severely damaged during wall retreat and subsequently causing gate end problems. Roofbolts had been used in the Sydney coalfield between 1949 and 1963 in room and pillar working, however, some concerns arose with using roofbolts as a primary support in long (2.2 - 3.4 km) single entry roadways. In order to alleviate these concerns, a phased approach was adopted whereby conventional steel densities remained along with roofbolts. The roofbolt design was such that the roofbolts would be the primary support and the steel sets would experience little or no loading. A detailed geotechnical monitoring program was developed by both CBDC and the Cape Breton Coal Research Laboratories technical staffs. The results of this monitoring were used to build-up confidence in roofbolts and subsequently reduce the steel set densities until eventually only roofbolts were used as a primary support. This paper will discuss briefly the phased introduction of rootbolts at Phalen Colliery and more specifically the development of geotechnical data collection and interpretation.

Jan 1, 1994

By K. E. Tame

Pollution from radioactive nuclides, principally Ra226, in liquid waste from uranium mills can be minimized by reusing the waste solution within the mill. Such reuse of waste solution in the Vitro Chemical Co.-type process was investigated in the laboratory. Vitro employs an acid leaching process for dissolution of the uranium, countercurrent decantation for liquid-solid separation and washing, and amine solvent extraction for recovery of the uranium from solution. The simulated reuse of barren raffinate in the mill circuit was carried through two separate laboratory campaigns of 29 and 31 cycles. Conventional chloride stripping of the amine was employed in the first campaign, whereas nitrate stripping was used in the second campaign. Dissolution of uranium from the are in the acid leaching circuit averaged 96 percent. Washing efficiencies of 98 to 99 percent were attained by using 4 volume tons of combined raffinate and makeup water per ton of ore. Uranium recovery from the pregnant solution using a 5-percent solution of amine LA-1 in kerosine ranged from 99.3 to 99.9 percent.

Jan 1, 1962

By S. Noll

Ventilation is an essential part of any underground mining operation. It provides fresh air for workers and machinery, keeps up an adequate mine climate and dilutes and dissipates noxious gases. Possible changes in the economic environment for mining in the future, e.g. increasing energy costs, will necessitate optimization of distribution of airflow within most mines. This paper discusses the scientific set-up for measurements as well as the behavior and influence of parameters of ventilation according to airflow distribution under consideration of technically possible set-ups of mine fans. This study was done in an underground mine with a flat seam room and pillar mining method.

Feb 27, 2013

By Christopher J. Bise

INTRODUCTION Traditionally, the purpose of mine ventilation has been defined as diluting, rendering harmless, and sweeping away hazardous gases which may accumulate in the mine workings. In addition, the ventilation sys- tem must keep the mine air as close as possible to the composition of the outside air in order to provide a healthy environment for mineworkers and to prolong the useful life of mine equipment. When plans are being developed for opening a new mine, sound engineering dictates that a calculated estimate be made of both the maximum air quantity requirement and the accept- able mine pressure (head) loss related to the under- ground layout and projections. For efficient mine ventilation, the two problem areas that demand constant attention are (1) the control of fugitive air (through short circuiting or leakage) and (2) the provision of ample airway capacity. Fugitive air is that quantity of air which enters the mine but short-circuits the mine ventilation system through improperly constructed stoppings, doors, overcasts, etc.; it represents an investment in power consumption (to move the air by the fan) while providing little, if any, return on the investment. Typically, efficient mines are those that can deliver 50% of the air handled by their fans to the working faces. Paying close attention to a new mine's ventilation system at the outset can provide numerous benefits during the later stages of mine development. It has often been noted that the main cause of trouble in many mine ventilation systems can be found within a radius of 2500 ft from the fan installation. Ample airway capacity is essential to maintain the mine's head loss within manageable limits. Similar to an industrial ventilation setting, the mine's headings are the ductwork for the handling of air. The difference between normal industrial ventilation and mine ventilation is apparent when more capacity is needed; a plant can install new or larger ductwork, but it is difficult for a mine to install new entries or enlarge existing entries. Air velocities are usually maintained under 800 fpm to limit the stirring up of dust in the intakes. With this in mind, it is usually simple to estimate the number of entries required in the ventilation system. The problem that often arises, though, is when the system's capacity is overestimated and the required head loss to achieve the desired quantity is excessive. Experience has shown that problems are often encountered when the mine's condition requires much more than three inches of water gage to ventilate the mine on a day-to-day basis. Thus, if proper attention is not paid to these parameters in the preplanning and initial development of a new mine, it may take an extremely high water gage to achieve the ultimate quantity. This chapter will not delve deeply into the theory of mine ventilation; many other books have been writ- ten on that subject. Instead, the following sections will point out the important considerations that should take place before the development of a mine ventilation system, as they relate to normal industry practice. FUNDAMENTALS OF AIRFLOW Volume Versus Water Gage The common method of measuring ventilation pressure is in terms of inches of water gage (wg); one inch of water corresponds to a pressure of 5.2 psf (pounds per square foot). Further, airflow follows a square-law relationship between volumes and pressures where, as shown in the previous section, a doubling of quantity requires a quadrupling of pressure. Table 1 shows how heads and quantities are related. Using Table 1, it is readily apparent that, for the same mining conditions, an increase in pressure from one inch wg to three inches wg results in only 34,640 cfm when the initial quantity is 20,000 cfm. Pressure Losses Pressure losses in mine ventilation systems are caused by friction between the heading walls and the air stream and by shock losses due to abrupt changes of the air stream caused by obstructions. The common pressure-loss formula for mine ventilation is: [ ] where H is the pressure loss in inches of water gage, K is the coefficient of friction, L is the length of the airway in feet, 0 is the airway perimeter in feet, V is the air velocity in feet per minute, and A is the cross-sectional area of the airway in square feet. Table 2 lists the US Bureau of Mines schedule of friction factors (K) for mine airways. Recent research indicates that the values of K are slightly less for coal mines than those presented in the USBM Schedule, and Table 3 should be consulted in this instance. It should be

Jan 1, 1986

By C. Allen

Vale Inco Ltd. is trialing several energy conservation initiatives in an effort to minimize costs and reduce green house gas emissions. The Ventilation on Demand (VOD) concept is a major project being carried out in various phases at Vale Inco mines. Due to the complex nature of a dynamic VOD system, Vale Inco is taking a systematic approach of testing VOD system components with milestone evaluations of each project. Coleman Mine was selected to pilot test the functionality and robustness of automated louvers in the underground environment; with the goal of optimizing mine airflow such that ventilation would be available for developing a new ore body. This project involved installing automated louvers in the exhaust raises; and installing variable frequency drive (VFD) starters on auxiliary supply fans. Because the VOD project created efficient utilization of the ventilation air in the 153 OB, the 170 OB is currently being developed without increasing the total fresh air into the mine. This paper describes the louver system component of VOD as applied at the mine.

Jan 1, 2009

By W. J. Kennedy, John D. Gardner

Introduction During the past 10 years, energy costs in mining have risen from about 4% to as much as 25% of total production cost. When combined with declining ore grades, these costs exert inordinate stress on mine productivity. Industrial engineering techniques have traditionally been aimed at improving manufacturing facilities productivity, but they have been applied to mining as pressure for productivity improvement has increased. These techniques can perform four useful functions for mining operations by improving methods, equipment selection procedures, energy use through scheduling, and managing energy costs rather than being managed by them. This article discusses some of the standard techniques for productivity improvement, and how they can be adapted to perform these four functions. Pre-Production The best time to seek high productivity is when a mine is being planned. During pre-production, when most options are still open, an optimum mine design is achievable. At this time, major equipment selections are made, the layout, material handling systems, and processes are designed, and critical energy source decisions are made. By including explicit energy cost projections, decisions can be made that will save significant amounts of money per ton of ore over the life of the mine. Attention can also be given to often overlooked details such as lighting. Some of the same tools used in new mine planning are applicable to productivity improvements at an existing plant. These tools are the charting and modeling methods that have been developed by industrial engineers over the past 75 years. Energy Consumption Before any analysis of energy consumption can be performed, preliminary work is necessary to find the kind of energy used, how much it costs the plant, and when electrical peaks occur. Past bills should be examined to determine the monthly consumption rate and the cost of each energy type. These data should be adjusted to remove effects caused by production fluctuations or weather. Knowing when electrical peak demands occur is necessary because these determine either the electrical capacity needed or the amount that must be bought. At $500-$800 per kW for construction or $5-$15 per kW per month for purchase, they represent a significant cost. By the time these data are collected, planners should be aware of the importance of energy in the total cost of mine operation, and should have the most crucial time data for electrical peak demand. Flow Process Chart Reducing costs can be accomplished by a thorough look at mine systems and how altering one system causes changes in others. By developing and exercising models of production processes outward to the transportation of the final product, otherwise obscured weaknesses may become apparent. There are tools available for analyzing the mine and its associated systems; one of the most useful is the flow process chart. The flow process chart gives a system analysis of a production procedure by classifying all actions into one of five types: operations, inspections, transportation, delays without paperwork, and delays with paperwork. Operation: In an operation, the nature of something is changed. Examples are blasting, washing coal, and processing matter. Denoted by 0. Inspection: Examining an object for identification or to measure some attribute. Examples are determining whether a drill hole is deep enough, deciding how far to cut under an unsupported roof when operating a continuous miner, or checking the lubricant level on a truck. Denoted bye. Transportation: Moving an object from one place to another. Examples are transporting overburden, moving coal on a conveyor belt, and pumping water. Denoted by =>. Delay without paperwork: A delay that is an inherent part of the system. Examples include idle time of a continuous miner while waiting for shuttle cars or for roof bolting, crushed ore in a storage area waiting for grinding and flotation, and people from one craft waiting for another craft to finish. Denoted by D. Delay with paperwork: A delay that can be terminated only by some written authorization. Examples are coal that cannot be removed until authorized by a bill of sale and transportation orders, personnel records, maintenance, or spare parts. Denoted by V. In practice, the flow process chart is completed after data collection and a detailed examination of the process. Data collection is needed to answer questions on distance, quantity, and time for each of the operations. As the analysis is being performed, each step of the process is subjected to these questions: • Why is the step necessary? • What is being moved? • When does the step take place, and why?

Jan 11, 1983

By David Allen

Cable bolting is a relatively new strata reinforcement technique that has been used successfully in Australia and the United Kingdom. Jim Walter Resources, Inc. has examined the principles and can visualize its useful application in many areas, such as: a) Longwall pull-out areas, b) Roadway junctions, particularly those involving yield pillars, c) Rib reinforcement, and d) Possible replacement of cribbing in roadways, especially longwall tailgates. This paper describes the conditions which led to cable bolt testing, the results of tests conducted to date, and conclusions about the direction that Jim Walter Resources, Inc. is heading as far as testing and potential applications are concerned.

Jan 1, 1994

By Don M. Coulter

"INTRODUCTION The Margerum bauxite district, in Colbert County, Ala. was investigated in December 1940 and in February 1943 by R. W. Smith of the Bureau of Mines .3/ A geologist of the Federal Geological Survey investigated and mapped the district in the fall of 1942.These preliminary investigations indicated bauxite showings over an area about a mile long by 2,000 feet wide. It was felt that the chances were good for finding one or more large deposits of low- to medium-grade bauxite. The location - within 35 miles of the alumina plant of the Reynolds Metals Co. atLister Hill, Ala. – made it especially attractive. A field office was established in Iuka, Miss., on May 19, 190,, and mineral agreements were executed with four landowners covering 1,146.6 cores.The drilling on the project was done under contract. The power drilling was supplemented by a labor crew that drilled hand-auger holes, where there was little or no gravel overburden, and sampled the old test pits.The Federal Geological Survey, after thorough study of the district, recommended 54 drill holes 4/ to delimit the known deposits and to search for possible extensions. After the drilling started, many of these locations eliminated, and others were changed upon the basis of information gathered from the holes as they were drilled."

Mar 1, 1948

By Oliver Bowles

ASBESTOS is a name applied to a group of naturally fibrous minerals. As indicated in a later section devoted to a detailed discussion of the early history of asbestos, the tern may be defined as a. fibrous form of serpentine or amphibole. It does not include fibrous forms of other minerals, such as wollastonite, nemalite (fibrous brucite), or the fibrous forms of calcite or gypsum. A material that has the characteristics of silk or cotton and at the same time will not burn is unique. Because it combines these qualities with other advantageous properties, asbestos has specialized uses for which no substitutes are yet available. Asbestos furnishes a major raw material for a great variety of essential products and their manufacture constitutes a vast industry. The United States leads all countries in the manufacture of asbestos products. The value of such commodities manufactured in 1952 exceeded $332 million. They are not only of vast importance to budding construction and industry but axe absolutely, essential to certain important fields of use, for example, steam packings and friction materials, such as brake-band linings on automobiles.

Jan 1, 1955

By M. R. Saharan

A brief overview of rockburst control measures is presented. The paper then elaborates on one key rockburst control measure, which is known as destress blasting. A historical background of the technique is presented in order to understand different aspects of its field application. The publication enumerates key aspects of the technique as being used in practice in South Africa and North America. The paper also presents an analysis of three case studies in order to demonstrate difficulties associated with the field application of distress blasting in hard rock mines.

Jan 1, 2006

By Michael Ostrowski

While the words “explosives” and “blasting” cause insurers to become apprehensive, skilled contractors using modern techniques under controlled conditions rarely cause serious damage or injury. However, blasting operations frequently result in complaints and, consequently, insurance claims. The purpose of this bulletin is to inform construction project participants of some of the measures which may be taken to reduce losses arising from blasting operations.

Jan 1, 2003

By L. Su, S. Liu, K. Jiang, K. Xie

The reduction of iron dissolution and acid consumption is fairly important in the extraction of Ni and Co from laterite ores from both economic and efficiency perspectives. BGRIMM conducted detailed research to investigate the behavior of iron in a nickel laterite leaching process. The mineralogy of Ni and Co was studied. In limonite ore about 76% of nickel is present in goethite while in saprolite ore 79% of nickel is associated with magnesium and silicon minerals. Test work revealed that limonite can be decomposed completely with enough acid concentration during atmospheric leaching. More than 98% of Ni and Co were leached into solution while 85% of the iron was also dissolved with 130-140 g/L iron in solution. Iron precipitates as hematite in the temperature range 250-270?C without neutralization reagents. That’s why limonite is treated by high pressure acid leaching (HPAL) technology to achieve nickel extraction and iron precipitation in the same autoclave. BGRIMM’s test results indicated that the remaining acid in atmospheric leaching solution and acid released by iron precipitation at a lower temperature was enough to leach saprolite efficiently. The so-called inverse leaching technology leaches limonite at atmospheric condition and the leaching slurry combined with saprolite is pumped into an autoclave to leach saprolite at 150-170?C. In the autoclave iron precipitates as hematite without acid consumption. Total nickel and cobalt extractions in the test program were more than 90% with iron content in solution below 8 g/L. In this inverse process iron dissolves in the atmospheric leaching stage and precipitates as hematite in the low temperature leaching stage. Compared with HPAL the leaching temperature is reduced to 150-170 ºC from 250-270 ºC. Compared to atmospheric tank leaching acid consumption is reduced from 850-950 kg/t ore (atmospheric) to 550-650 kg/t ore (inverse).

Jan 1, 2016

By A. Jahedsaravani, M. Massinaei, M. Zarie

Deep learning is a subset of machine learning that uses artificial neural networks for extracting high-level features from image data. In the present study, a soft sensor is proposed for the prediction of the flotation performance through froth features generated by the use of pre-trained convolutional neural networks. Several state-of-the-art convolutional neural networks (AlexNet, GoogLeNet, VGGNet, ResNet, and SqueezeNet) pre-trained on the ImageNet database are used to predict the metallurgical performance of two flotation systems. The first case study is a batch copper flotation system video-captured over a wide range of process conditions. The second case study is an industrial coal flotation column equipped with a continuous video recording system. The pre-trained networks are used to extract features from the froth images, and these features are subsequently used to predict the flotation conditions and performance. The prediction results by the pre-trained algorithms were compared with the traditional image processing algorithms. This demonstrates the ability of the pre-trained structures to generalize to images outside the ImageNet database. GoogLeNet outperforms other network architectures and provides more accurate predictions of the flotation process behavior and performance.

May 5, 2023