Search Documents

By F. Awotwe-Appiah, P. Koshy, A. Y. Fosu, J. Dankwah, J. R. Dankwah

Ghana currently exports its bauxite raw without any value addition. Gradual depletion of high grade bauxite (with high alumina content) and the emergence of low grade bauxite (with high iron oxide content) calls for novel and inexpensive ways to recover alumina from these low grade ores. Bauxite users have indicated their willingness to pay 10% to 50% higher for a corresponding reduction in impurities by a similar amount. This project therefore investigates the carbothermal upgrading of the Ghana Awaso bauxite using waste polypropylene as reductant in a locally designed gas-fired furnace. Composite pellets of low grade bauxite sample (assaying 45% Al2O3) containing pulverised waste polypropylene were formed, oven dried, placed in fireclay crucible and fired for 30 minutes. The magnetic fraction after reduction was separated by a Low Intensity Magnet and weighed. It was observed that up to 97% of the iron oxides (occurring as hematite and goethite) could be separated into the magnetic fraction as magnetite leaving a nonmagnetic fraction with over 98% Al2O3 as opposed to the 45% Al2O3 in the as-received bauxite sample. This research direction provides yet another simple but effective way for simultaneously upgrading low grade bauxite and recycling waste polypropylene, whilst at the same time decreasing the amount of waste material generated as red mud after the Bayer process.

Jan 1, 2015

By Behrooz Abbasi, Seyedsaeid Ahmadvand, Thom Seal, Babak Azarfar, Bryan Ulrich

Hydro-Jex® is a new enhanced method for heap leach treatment with significantly higher mineral recovery. In this case study at the Los Filos Mine in Mexico, the Hydro-Jex operation is evaluated for heap leach stability. This is pursued by numerical modeling of the heap leach pad stability and well monitoring of the phreatic surface before and after the Hydro-Jex operation. Experimental and numerical results both indicate an improvement in mineral processing under the influence of Hydro-Jex. Numerical results show an insignificant decrease and a significant increase in stability during and after Hydro-Jex injection application, respectively. The integrity of the liner is maintained in both cases. The zonal injection pressure, operationally below the overburden pressure, is found to be negligible compared with the dimensionality of the heap, and stability of the structure is not exposed to any significant risk, even up to 25%above overburden pressure. Furthermore, heap leach factor of safety increases after Hydro-Jex application. This is attributed to breakdown of the water solution build-up and thus a decrease in the phreatic surface depth below the top of the pad. It is suggested that injection wells be drilled, based on thorough geophysical data, in locations where overcompaction of heap material results in water solution build-up and pore pressure enhancement. Compared with traditional heap leaching, the Hydro-Jex technique not only expedites mineral processing by increasing chemical kinetics extraction but also increases the stability of the heap by unclogging drains.

Jun 29, 2020

By K. C. Dean

During investigation of the occurrence of manganese deposits in the western United States, examining engineers of the Bureau of Mines collected samples from several properties in Montana. Extensive laboratory tests were made to determine the amenability of these ores to standard ore-dressing methods of concentration to produce marketable-grade products. The Tzarena claim, one of the properties sampled, is on the western end of the Rama or Black Chief vein in the city of Butte, Silver Bow County, Mont., on the southwestern periphery of the mining district. The Black Chief 'WAR system contains one of the largest and most extensive known rhodochrosite deposits in the world. Test work was conducted on two samples of Tzarena ore. Sample 1 is representative of a vain 5 feet wide on the 150-foot level of the Tzarena shaft. Sample 2 is representative of a vein 8 feet wide in the 100-foot level of the shaft. The Minnie Jane mine, 1/2 mile northwest of the Tzarena claim, contains ore similar in mineralogical composition to Tzarena ore but differing materially in manganese content and to some extent in physical characteristics. Ore-dressing investigations were made on three samples of Minnie Jane ore. Sample 1 represented stock-piled ore, sample 2 represented ore from Minnie Jane mine having a considerable sulfide content, and sample 3 represented low-grade ore.

Jan 1, 1951

By Hossein Ghaedi, Hadi Fattahi

Despite the significant role of coal in the economic progress of nations, the occupational and health risks associated with its mining pose a major concern for industry stakeholders. The occurrence of roof collapses in coal mines remains a critical factor leading to substantial loss of life and financial damages for miners. Therefore, accurately predicting the roof fall rate (RFR) holds paramount importance. However, the uncertainty surrounding rock parameters in mines hinders the application of conventional methods to assess roof collapse rates in coal mines. To tackle the challenges associated with predicting roof fall rates in underground coal mines, this study proposes a novel solution by leveraging the Rock Engineering System (RES) method. The investigation is grounded in a dataset comprising 109 data points, encompassing crucial input parameters like depth of cover (DOF), primary roof support (PRSUP), intersection diagonal span (IS), mining height (MH), and coal mine roof rating (CMRR). In the model construction phase, 80% of the data (87 points) were utilized to build the RES model. A critical aspect of this study involves the evaluation of the RES model’s performance against alternative regression techniques, namely linear, power, exponential, polynomial, and logarithmic regression. This comparison was executed using the remaining 24 data points (20% of the dataset) for rigorous evaluation. Employing key statistical metrics such as mean square error (MSE), root mean square error (RMSE), and squared correlation coefficient (R2), the study systematically demonstrated the superior accuracy of the RES-based method compared to other approaches. In conclusion, the outcomes strongly support the efficacy of the RES method in predicting roof fall rates, not only in the specific case studied but also indicating promise for its application in other underground coal projects. This underscores the potential of the RES method as a reliable and versatile tool for forecasting roof fall rates in the complex and critical context of underground coal mining.

Mar 19, 2024

By M Z. Abzalov

Drilling mineral sand deposits has specific challenges, which are different from drilling hard rock deposits: Drilled sequences are represented by non-consolidated free-flowing sands to semi-consolidated clayey sands and indurated layers. Valuable heavy minerals (ilmenite, rutile, zircon, leucoxene) have an average density of 4.6 g/cm3, which is significantly heavier than the hosting sand matrix; mainly represented by quartz (density = 2.6 g/cm3). The large difference in the mineral densities can result in sample segregation in the drilling rods and capturing devices causing biased composition of the recovered samples. The drilled strata can contain beds of consolidated sediments, hard pan lenses and concretions. Therefore, chosen drilling equipment should be flexible and permit a switch from drilling the free-flowing sands to hard rock drilling. Thickness of the sand hosted mineralisation varies from several metres to more than 170 m. Mineralisation is distributed above and below the water table. Evaluation of the mineral sands deposits is based on hundreds of drill holes and many thousands of samples, therefore drilling methods should be fast and cost effective. The drilling methods commonly used for mineral sand deposit evaluation, include aircore, triple tube diamond and sonic; including its variants, such as vibrocore drilling technique. The ability of the studied drilling techniques to properly address the above listed challenges in mineral sand deposits, whilst maximising the quality of the obtained samples has been compared in three deposits: Corridor Sands (Mozambique), Richards Bay (Republic of South Africa) and Fort-Dauphin (Madagascar).This study has shown that aircore samples tend to be negatively grade biased, underestimating the heavy minerals content, in particular when drilled intervals are located below the water table. Triple tube diamond core method can produce non-biased samples. However, results can be marred by incorrectly chosen drilling mud. Sonic drilling is a relatively new technique based on a high frequency vibratory technology. It is the only production drilling technique that allows a continuous relatively undisturbed sample of unlithified sands to be obtained. The main limitations of this technology preventing its wider use for developing mineral sand resources are the relatively high drilling costs, in particular when the depth of drilling exceeds 100 m, and speed of drilling advance.

Oct 5, 2011

By L Casari

A study of samples from three sectors of the mine, which was worked from Roman times to 1991, revealed different textures, mineral associations and distributions of trace elements in the sulphides. In the last decade before closure, grades of sphalerite concentrate were in the range 51.67-53.30% Zn and 2.21-2.86% Pb and of galena concentrate 69.43-71.62% Pb and 4.70-8.36% Zn, while the estimated by-product germanium content of sphalerite concentrates sold was 300-400 g/t. Passing the different run-of-mine materials through the same concentrator flowsheet produced middlings and tailings during the preconcentration processing. Separate and alternative treatment of run-of-mine products would have yielded improved recoveries of sphalerite and galena and better values of by-product germanium and cadmium

Jun 18, 1905

By Paavo Eerola

In May of 1984, it was decided that Outokumpu Oy would design and build a new nickel flotation mill to process ore from the newly discovered Enonkoski mine in eastern Finland. Since there hart peen a number of successful process charges in the existing mills, the design effort included the study and incorporation of several new developments, including all-autogenous grinding, "flash flotation", lane cell roughing, large cell cleaning and a computer-based, profit optimizing, process control system. The paper discusses the background and results of those studies, and a description of the resulting mill which will begin production in 1986.

Jan 1, 1985

By J Nell, J Petersen, X C. Goso

The feasibility of increasing the deportment of titanium species and the stability of the pseudobrookite solid solution (M3O5) in a titaniferous slag was investigated, to facilitate the valourisation of titaniferous slags. Thermochemical simulations using FactSage™ software, ver 8.3 (by Thermofact/CRTC and GTT-Technologies), and equilibration-quench experiments were conducted using an appropriate synthetic titaniferous slag composition in an oxygen partial pressure (pO2) ranging from air to 10-16 atm under non-isothermal conditions. The temperature-pressure phase stability diagram for the synthetic slag containing 37.19 per cent TiO2, 19.69 per cent SiO2, 13.12 per cent Al2O3, 16.00 per cent CaO and 14.00 per cent MgO showed that the spinel solid solution (M3O4) was the primary phase until a pO2 of 10-10.5 atm, thereafter M3O5 was the primary phase until a pO2 10-13.2 atm. M3O4 again was the primary phase between a pO2 10-13.2 and 10-16 atm. The changeover of the primary phase coincided with an increase in the liquidus surface. When cooling the slag in a pO2 of 10-16 atm, M3O5 crystallised as a ternary phase that showed a rapid increase in mass while that of the primary M3O4 phase decreased and that of the secondary perovskite (CaTiO3) phase increased at a slower rate. At 1300°C under sub-liquidus conditions, the contents of M3O5, M3O4 and CaTiO3 were about 20 per cent, 15 per cent, and 11 per cent, respectively. Experimental results showed that the Ti3+/Ti4+ mass ratio in the M3O5 phase, ie a measure of the reduction of titanium species, is not considerably sensitive to equilibration temperature in the reviewed range. Slow cooling and minimum/no oxidation during cooling of titaniferous slag would increase the deportment of titanium species to, and the growth and stability of the M3O5 phase.

Jun 19, 2024

By J. Esson, R. A. D. Pattrick, M. K. Nimfopoulos, D. A. Polya, K. M. Michailidis

"Economic Mn-oxide ore deposits of commercial grade occur in the Rhodope massif near Kato Nevrokopi in the Drama region, Northern Greece. The Mn-oxide mineralization has developed by weathering of continental hypogene rhodochrosite-sulphide veins. The vein mineralization is con-?ned by tectonic shear zones between marble and metapelites, extending laterally into the marble astabular, pod or lenticular oreshoots (up to 50 m x 20 m x 5-10 m). Supergene oxidation of the hypogene mineralization led to the formation of in-situ residual Mn-oxide ore deposits, and secondary infills of Mn-oxide ore in embryonic and well developed karst cavities.Whole rock geochemical pro?les across mineralized zones con?rm the role of thrusts and faults as solution passageways and stress the importance of these structures in the development of hydrothermal and supergene mineralization at Kato Nevrokopi. Three zones are recognized in the in-situ supergene veins: (A) a stable zone of oxidation, where immobile elements form (or substitute in) stable oxide mineral phases, and mobile elements are leached; (B) a transitional (active) zone in which element behavior is strongly in?uenced by seasonal ?uctuations of the groundwater table and variations in pH-Eh conditions; and (C) a zone of permanent ?ooding, where variations in pH-Eh conditions are minimal. Zone (B) is considered as the source zone for the karst cavity mineralization. During weathering, meteoric waters, which were CO2-rich (PCO2~10-3.8 to 10-1.4) and oxygenated (fO2~10-17 for malachite), percolated downward within the veins, causing breakdown and dissolution of sul?des and marble, and oxidation of rhodochrosite to Mn-oxides. Karst cavity formation was favored by the high permeability along thrust zones. Dissolved Mn2+was transportedinto karst cavities in reduced meteoric waters at the beginning of weathering (pH~4-5), and as Mn(HCO3)2in slightly alkaline groundwaters during advanced weathering (pH~6-8). Mn4+-oxide precipitation took place by fO2increase in ground waters, or pH increase by continuous hydrolysis and carbonate dissolution. In the well developed karst setting, some mobility of elements occurred during and after karst ore formation in the order Na>K>Mg>Sr>Mn>As>Zn>Ba>Al>Fe>Cu>Cd>Pb."

Jan 1, 1997

By J. Esson, R. A. D. Pattrick, M. K. Nimfopoulos, D. A. Polya, K. M. Michailidis

"Abstract-Economic Mn-oxide ore deposits of commercial grade occur in the Rhodope massif near Kato Nevrokopi in the Drama region, Northern Greece. The Mn-oxide mineralization has developed by weathering of continental hypogene rhodochrosite-sulphide veins. The vein mineralization is confined by tectonic shear zones between marble and metapelites, extending laterally into the marble as tabular, pod or lenticular oreshoots (up to 50 m ?? 20 m ?? 5-10 m). Supergene oxidation of the hypogene mineralization led to the formation of in-situ residual Mn-oxide ore deposits, and secondary infills of Mn-oxide ore in embryonic and well developed karst cavities.Whole rock geochemical profiles across mineralized zones confirm the role of thrusts and faults as solution passageways and stress the importance of these structures in the development of hydrothermal and supergene mineralization at Kato Nevrokopi. Three zones are recognized in the insitu supergene veins: (A) a stable zone of oxidation, where immobile elements form (or substitute in) stable oxide mineral phases, and mobile elements are leached; (B) a transitional (active) zone in which element behavior is strongly influenced by seasonal fluctuations of the groundwater table and variations in pH-Eh conditions; and (C) a zone of permanent flooding, where variations in pH-Eh conditions are minimal. Zone (B) is considered as the source zone for the karst cavity mineralization. During weathering, meteoric waters, which were CO2-rich (PCO2 ~10-3.8 to 10-1.4) and oxygenated (fO2 ~10-17 for malachite), percolated downward within the veins, causing breakdown and dissolution of sulfides and marble, and oxidation of rhodochrosite to Mn-oxides. Karst cavity formation was favored by the high permeability along thrust zones. Dissolved Mn2+ was transported into karst cavities in reduced meteoric waters at the beginning of weathering (pH~4-5), and as Mn(HCO3)2 in slightly alkaline groundwaters during advanced weathering (pH~6-8). Mn4+-oxide precipitation took place by fO2 increase in ground waters, or pH increase by continuous hydrolysis and carbonate dissolution. In the well developed karst setting, some mobility of elements occurred during and after karst ore formation in the order Na>K>Mg>Sr>Mn>As>Zn>Ba>Al>Fe>Cu>Cd>Pb."

Jan 1, 1997

By W. Dingley, G. R. Hoey

"AbstractWear of six types of grinding balls was compared with and without corrosion inhibition in grinding Ni-Cu sulphide and hematite ores in laboratory ball mills. Forged steel balls (Rc 60 to 65) and three types of white cast iron balls (Rc 50 to 65) were susceptible to corrosive wear and two types of stainless steel balls [AISI Type 440C (Rc 65) and AISI Type 302 is; 35)] were relatively immune. The relative wear rates of the five types of hard balls (Rc 50 to 65) were dependent on the corrosivity of the ore. Ranking of the hard balls with respect to wear resistance was consistent for the sulphide and hematite ores in the presence of corrosion inhibition, whereas their ranking was different for the two ores in the absence of inhibition.The electrochemical potential of the grinding balls was measured during grinding and was found to be a useful measurement for assessing the susceptibility of an alloy to corrosive wear. Reduction of ball wear for forged steel and white cast irons with corrosion inhibition was concomitant with a shift of ball potential to less active values. Ball potentials in the presence of inhibition were indicative of non-corroding metal surfaces.IntroductionA PREVIOUS STUDY by the authors (1) showed that corrosion inhibitors - sodium nitrite, sodium chromate and sodium metasilicate - reduced wear of forged steel grinding balls by up to 49% in grinding Ni-Cu sulphide ore. Although the mechanism of wear reduction was not elucidated, it was concluded that corrosion must play a significant role in the wear of grinding media. Remark and Wick (2) of Battelle Pacific Northwest Laboratories have also evaluated corrosion inhibitors for the reduction of metal wear using abrasive and n on- abrasive te sting procedures. Their research was sponsored by 16 companies throughout the world. Preliminary results indicated that anodic inhibitors prevent from one-fourth to one-half of the loss due to corrosion."

Jan 1, 1977

By Enrique Roy Dionisio Calderon

O presente trabalho mostra a concepção de um modelo que permite gerar uma superfície de mínimo consumo de redutor (carbono) em Altos-Fornos. No seu desenvolvimento foram consideradas as condições de equilíbrio das reações de redução, os balanços de massa e de energia e parâmetros operacionais. Sua aplicação à redução em Altos Fornos é exemplificada para situações reais, baseada em banco de dados operacionais de uma usina siderúrgica. Foi possível, a partir de planos de corte nestas superfícies, gerar diagramas binários do tipo CR-RD (Consumo de Redutor-Fração de Redução Direta), com possibilidade de serem gerados “on-line” no processo. Dentre os resultados obtidos, destaca-se a identificação de inviabilidades termodinâmicas importantes do forno. Dentre elas o modelo permitiu, através de seus “outputs” gráficos, prever e acompanhar o fenômeno de “gelada de forno”, conferindo desta forma maior segurança operacional ao processo.

Aug 17, 2017

Cutter roof failure, a ground control problem that frequently results in massive roof failure, is common in coal mines of the Northern Appala-chian Coal Basin, causing delays in production and posing a safety hazard to mine personnel. The Bureau of Mines is conducting research on the causes of cutter roof failure to gain a basis from which to predict and prevent its occurrence, and to support such roof when failure occurs or is imminent. Research conducted at the Greenwich Collieries North and South Mines, in central Pennsylvania revealed a correlation between the presence of clastic dikes and formation of cutter roof failure. The research con-sisted of in-mine mapping of geologic and deformational features, rock pressure monitoring, core recovery and evaluation, and in situ stress measurement. It was found that roof failure increased in areas where the occurrence of clastic dikes was most frequent, especially where two or more clastic dikes intersect. Staggering of crosscuts prevented the extension of cutters and deterred the occurrence of large falls extending over several break-throughs. In addition, trusses and cribbing were effective in stabilizing clastic dikes and inhibiting the development of cutter roof failure when employed immediately or shortly after mining.

Jan 1, 1987

By M. F. Goodrich

The concept of using water-only cyclones as a clean coal scalping device preceding a heavy media cyclone circuit shows considerable promise, as it can considerably reduce the feed to the heavy media cyclone circuit. The system will not perform as well as a heavy media cyclone circuit alone, but the resultant loss in yield is small and may be negligible under some conditions. The concept of water-only cyclone scalping has previously been suggested by others. By reducing the amount of raw coal to be processed in the more expensive circuits overall capital and operating costs may be reduced. There are a few commercial plants with circuits operating on this concept. In the Warwick and Rowland plants water-only cyclones are used to scalp the feed to a heavy media vessel and Deister tables respectively while Amherst Coal Company's McGregor plant uses them to scalp the feed to a heavy media cyclone circuit. The system was evaluated on a technical basis, both from a theoretical viewpoint and by analysis of a computer simulation, in order to gain a fuller understanding of its usefulness and limitations.

Jan 1, 1977

By T. M. Barczak

The term "resultant load vector" is defined as the representation of the forces applied to a longwall roof support element by strata activity through a single, quantifiable measure of support resistance. The relatively complex kinematics of the shield structure prohibits determination of support resistance simply from the summation of leg forces. In the research reported in this Bureau of Mines study, the mechanics of the shield structure were evaluated, and a technique was developed whereby the resultant shield loading could be determined by instrumenting supports with pressure transducers and strain gauges to measure leg, canopy capsule, and lemniscate link forces. This technique has been laboratory tested in the Bureau's Mine Roof Simulator. Functional relationships among variables were assessed, and confidence intervals were established for prediction of the resultant load vector parameters. Resultant load measurements were taken on five instrumented shields on an active longwall face in Colorado. Results were analyzed and found to be consistent with shield mechanics and anticipated roof behavior. Benefits to be derived from this research, future efforts, and long-range goals are discussed.

Jan 1, 1986

By Afshin Sadri

Uma das principais razões para o reparo de altos-fornos é a perda do revestimento refratário, principalmente nas regiões da soleira e onde são feitos os furos de corrida. A operação intensa dos altos-fornos pode resultar em desgaste e deteorização extensiva do material refratário. Isto pode, então, afetar a continuidade das operações e a vida útil do forno. Novo-Lipetsk (NLMK) e a Hatch desenvolveram uma abordagem sistemática para monitorar a condição do revestimento da soleira dos altos-fornos usando Acousto Ultrasonic-Eco (AU-E), um sistema de ensaios não-destrutivos desenvolvido pela Hatch. Inspeções contínuas e periódicas dos altos-fornos revelaram áreas com deterioração acelerada do material refratário, formação extensa de “elefant foot” e acelerado desgaste do refratário, assim como fissuras e outras anomalias. Para prolongar a vida do forno, mantendo a intensidade de suas operações, algumas modificações, tais como a melhoria da qualidade do coque e da limpeza das placas de resfriamento e adição de óxido de titânio, entre outras medidas, foram recomendadas e aplicadas. Este trabalho demonstra como esforços coordenados entre a NLMK e a Hatch resultaram no prolongamento da vida útil de um alto-forno

Aug 17, 2017

By Theo. A. Link

ACCORDING to our office dictionary, geology is "the science that investigates the structure of the earth and its successive physical changes", and a geologist is "one who knows the science of the structure of the earth". No doubt this definition does not suit all of us. Another dictionary would have the definition slightly different. As an example, one dictionary stated that "a geologist is one who practices geology". (Some of our clients told us to quit practising and get to work). There are many different types of geologists engaged in various branches of the profession and some of these are further broken down into subdivisions so that specialization is becoming more and more prevalent. As a result it is, in some cases, doubtful whether or not we are actually dealing with a geologist in the generally accepted sense of the term.

Jan 1, 1956

By K. E. Tame

A resume of the history and state of the art of physical beneficiation of uranium ores is presented along with a detailed description of research on concentration of uranium ores by attrition grinding. The information on flotation, gravity concentration, electronic sorting, and selective grinding was gathered to complement the study of attrition grinding methods. Both dry and wet grinding procedures were investigated as potential methods of upgrading a variety of ore types, ranging in grade from 0.02 to 0.42 percent U308. A Mardun4 disintegrator that employs high-speed rotating paddles in a cylindrical shell lined with interrupter bars was used in most of the dry grinding studies. Attrition scrubbing devices, which impart a violent mixing or attritioning action to ore pulps, and cylindrical mills, employing rods or punchings as grinding media, were used in the wet grinding tests.

Jan 1, 1961

By Thomas J. Peters

The Danskin/South Fork Boise River study area is 40 air km southeast of Boise, Idaho. The 16,000-hectare, two-segment area is mainly underlain by Cretaceous-age granodiorite and an aplitic and pegmatitic complex of the Idaho batholith, Miocene-age rhyolitic tuffs and lavas, and late Tertiary-age basalt flows. No mineral resources were identified in the study area. One mine, The Flat Creek mine, which produced several ounces of gold and silver, was found in the study area. It is geologically similar to deposits in the Neal mining district, adjacent to the northwest corner of the area; gold mineralization occurs along east-northeast-trending faults and associated quartz veins, which dip moderately to steeply southeast, and within adjacent Cretaceous granitic rock. Samples indicate the Flat Creek deposit is not rich enough to mine underground, and near surface, open-pit-mineable deposit areas could not be identified. However, exploration is warranted there and also in the 7-km interval to mines in the Neal district, and for extensions of those mines. No evidence of geothermal activity was observed, although there are hot springs in the vicinity. Some alluvial (placer) samples from the interior of the study area contained free gold, but no substantial gravel deposits were seen.

Jan 1, 1993

By Y. Fu, C. Aldrich

This study investigates the use of dense convolutional neural networks to estimate bubble size distributions in flotation froths. 20 videographically captured images of froths from an industrial platinum flotation plant in South Africa were labelled semantically with an image annotation tool, LabelMe (https://github.com/wkentaro/labelme). As a result, pixels in the images could be categorized into three classes, namely ‘bubbles’ (class 0), ‘bubble boundaries’ (class 1) and ‘background’ (class 2). A U-net architecture with a ResNet34 backbone was trained to segment the images via classification of the pixels in the images. 5 of the 20 images were used to validate the performance of the neural network. Explicit inclusion of the bubble boundaries as a class proved to be significantly more effective than training the network with two classes only (i.e. bubbles and background). However, the thickness of the boundary layer regions were critical. If too thin, it was difficult to distinguish between some contiguous bubbles and if too thick, some fine bubbles may not be identified. The results were promising, although there is still significant scope for improvement, among other by making use of more images for training and validation. Keywords: Flotation Froths, Convolutional Neural Networks, Deep Learning, Bubble size Analysis

Jan 1, 2020