Some aspects of ferrohydrostatic separation of minerals and the recycling of ferrofluid

• Main Author: Dumbu, Stanford
• Other Authors: Svoboda, J.
• Format: Others
• Language: en_ZA
• Published: Stellenbosch : Stellenbosch University 2012
• Subjects: Separation (Technology); Metallurgy; Dissertations > Chemical engineering; Theses > Chemical engineering
• Online Access: http://hdl.handle.net/10019.1/52456

Thesis (MScEng)--University of Stellenbosch, 2001. === ENGLISH ABSTRACT: Ferrohydrostatic separation (FHS) of materials is a float and sink technique which utilizes ferrofluid exposed to a non-homogeneous magnetic field. The efficiency of material separation depends on numerous variables. The most important variables, which were investigated individually, are the effects of moisture content, ferrofluid level, feedrate, particle size and material density distribution on separation efficiency. It is important to recover and recycle the ferrofluid attached to the products of separation so as to reduce the cost of the FHS technology and the amount of kerosene disposed of to the environment. This prompted research into some of the factors affecting ferrofluid recovery. The factors that were investigated are the effects ofFHS operation, material moisture content, particle size and porosity. The separation efficiency was found to be dependent on all the variables investigated. The effect of material moisture content is less pronounced for particles larger than 2.8 mm. This implies that wet feed material should be screened before ferrohydrostatic separation and material which particle size is less than 2.8 mm should preferably be treated dry. Wet material (less than 2.8 mm) floats, even though its density is greater than the cut-point density. This is owing to the immiscibility of the water coating the particles and the kerosene-based ferrofluid used for separation. It was found that the effect of ferrofluid level on separation efficiency is a function of both the density difference of the particles to be separated and the particle size. Separation efficiency as a function of ferrofluid level is poor for particles larger than 2 mm and is good when the density difference of the material to be separated is high, for instance between 2700 kg/nr' and 3530 kg/nr'. This shows that for efficient separation of coarse particles and near density material (material with density close to the cut-point density), the ferrofluid level should be controlled, preferably close to the maximum possible level. The effect of feedrate on separation efficiency is also a function of the densities of the particles to be separated. An increase in feedrate leads to poor separation for particles with densities close t~ each other. This implies that separation of near density material requires accurate feedrate control. It has been shown from simulation and modelling that the effective cutpoint density changes as the particle moves through the chamber until it eventually reaches its terminal velocity, given that the chamber is of sufficient size for this to occur. The effective cut-point density increases to the maximum as the particle enters the ferrofluid pool but settles down to a relative constant once the particle has reached its terminal velocity. The effective cut-point density was shown to decrease with an increase in particle magnetisation. It was found that this decrease in the cut-point density determines the density difference (difference between two particles) achievable when non-magnetic material is treated together with magnetic material. It is therefore important to magnetically scalp the feed material for efficient separation. When the material is not scalped, magnetic and nonmagnetic material with the same density might report to different density fractions, which leads to poor separation. This magnetic contribution to the effective density can be utilised in the separation of material with same density but different magnetisation. The efficiency offerrofluid recovery was found to be dependent on all the variables investigated. The amount of ferrofluid drawn from the FHS separator was found to decrease with an increase in the magnetic field. Furthermore, the amount of ferrofluid that remains attached to the particles after allowing ferrofluid to drain from the material is the same as that attached to the FHS products of separation at high magnetic fields. This shows that it is important to operate the ferrohydrostatic separator at high magnetic fields in order to attract most of the ferrofluid back to the separator. T-heamount of ferrofluid adsorbed onto and absorbed by the particles was found to decrease with an increase in the material moisture content. This is due to two factors. The first is that water occupies the vacant pores in the material. The second is that water forms a layer on the particle surface which is immiscible with kerosene-based ferrofluid. This phenomenon leads to a reduction in cost of the ferrohydrostatic separation technology when wet material as opposed to dry material is treated. As already described coarse material larger than 2.8 mm can be treated wet without detrimental effects on separation. For -8+4 mm particles, the ferrofluid loss ranges from 0.6 down to 0.14 kg/tonne of feed for 0 to 10 % material moisture content respectively. The amount of ferrofluid lost per tonne of feed was found to range from 0.73 to 0.56 kg for-O.85+O.5 mm to -12+8 mm particle sizes respectively. The increase in ferrofluid loss in small particles is due to the increase in surface area in small particles for ferrofluid adsorption. The increase in porosity increases the amount of ferrofluid lost due to the difficulties in recovering ferrofluid embedded in the pores of the particles. Adding water to coarse material lowers the amount of ferrofluid lost by reducing porosity. Modelling the amount of ferrofluid lost, as a function of particle size and porosity, would assist in determining the amount of ferrofluid required to treat a known amount of material. The quality of ferrofluid recovered was found to be the same as that initially used for material separation. This implies that the separation efficiency would not be affected by the use of recycled ferrofluid. === AFRIKAANSE OPSOMMING: Ferro-hidrostatiese skeiding van materiale is 'n flotasie (dryf) en besinkingstegniek wat gebruik maak van ferro-vloeistof wat blootgestel is aan 'n magnetiese veld. Die effektiwiteit van die materiaal skeiding is afhanklik van verskeie veranderlikes. Die belangrikste veranderlikes wat die skeidingseffektiwiteit beïnvloed is individueel bestudeer, naamlik voginhoud, ferro-vloeistof vlak, voertempo, partikelgrootte en materiaal digtheid verspreiding. Dit is belangrik om die ferro-vloeistof te herwin en te hergebruik om die koste van die proses en tegnologie te verminder en dus ook die hoeveelheid keroseen wat aan die omgewing blootgestel is. Dit het navorsing tot gevolg gehad oor die faktore wat ferro-vloeistofherwinning beïnvloed. Hierdie.faktore wat ondersoek is in hierdie studie is materiaal voginhoud, partikelgrootte en porositeit. Die skeidingseffektiwiteit was afhanklik van al die faktore wat ondersoek is. Die effek van materiaal voginhoud was minder beduidend vir partikels groter as 2.8 mm. Dit wys dat nat voermateriaal moet gesif word voor ferro-hidrostatiese skeiding, en materiaal met 'n partikelgrootte kleiner as 2.8 mm moet verkieslik gedroog word. Nat materiaal (minder as 2.8 mm) floteer selfs as die digtheid groter is as die snypunt digtheid. Dit is as gevolg van ondeurlaatbaarheid van die water wat die partikels bedek en die keroseen basis ferro-vloeistofwat gebruik word vir die skeiding. Dit is gevind dat die invloed op die skeidingseffektiwiteit van die ferrovloeistof vlak is 'n funksie van beide die digtheid van die partikels wat geskei word. Die partikelgrootte skeidingseffektiwiteit as 'n funksie van die ferro-vloeistof vlak, is swak vir partikels groter as 2 mm en is goed wanneer die digtheid verskil van die materiaal wat geskei moet word, hoog is, byvoorbeeld 2 700 kg/nr' en 3 530 kg/nr'. Dit wys dat vir die effektiewe skeidings vir groter partikels en naby digtheid materiaal (materiaal net 'n digtheid nabyaan die snypunt digtheid), moet die ferro- . vloeistofvlak baie goed beheer word, gewoonlik naby die maksimum vlak moontlik. Die effek van voertempo op die effektitiwiteit van skeiding is ook 'n funksie van die digtheid van die partikels wat geskei moet word. 'n Toename in die vloeitempo lei tot 'n swak skeiding van partikels met digthede wat naby mekaar lê. Dit wys weer daarop dat die skeiding van naby digtheid materiaal het die akkurate beheer van voertempo tot gevolg. Dit is gevind deur simulasie en modulering dat die effektiewe snypuntdigtheid verander soos die partikel deur die kamer beweeg totdat dit uiteindelik sy finale snelheid bereik (gegee dat die kamer groot genoeg is). Dit effektiewe snypunt digtheid verhoog tot 'n maksimum wanneer die partikel die ferro-vloeistof binne gaan, maar bereik na 'n kort tydperk 'n kostante waarde sodra die partikel sy finale snelheid bereik het. Die effektiewe snypunt digtheid verlaag met 'n toename in partikel magnetisme. Dit is gevind dat die afname in die snypunt digtheid bepaal die digtheidsverskil (verskil tussen twee partikels) wat bereikbaar is wanneer nie-magnetiese materiaal saam met magnetiese materiaal behandel word. Dit is dus belangrik om die voer materiaal magneties te skalpeer vir effektiewe skeiding. Wanneer die materiaal so behandel word, sal magnetiese en nie-magnetiese materiaal, met die dieselfde digthede, rapporteer in verskillende digtheidsfraksies wat sal lei tot swak skeiding. Die magnetiese bydrae tot die effektiewe digtheid kan gebruik word in die skeiding van materiaal met dieselfde digtheid, maar met verskillende magnetismes. Die effektiwiteit van ferro-vloeistof herwinning is afhanklik van al die veranderlikes wat ondersoek is. Die hoeveelheid ferro-vloeistof wat omtrek is van die ferro-hidrostatiese skeier verminder met 'n toename in die magnetiese veld. Verder is die ferro-vloeistof wat agterbly as gevolg van die feit dat hulle vas is aan die partikels na dreinering, dieselfde as die hoeveelheid wat vasgeheg is aan die die ferro-hidrostatisiese produkte van skeiding by hoë magnetiese velde. Dus is dit belangrik om die ferrohidrostatiese skeier te bedryf by hoë magnetiese velde om sodoende die meerderheid van die ferro-vloeistof in die skeier agter te laat bly. Die hoeveelheid ferro-vloeistof geadsorbeer aan en geabsorbeer deur die partikels verlaag met 'n toename in die materiaal voginhoud. Dit is gevolg van twee redes, nl. eerstens water wat die plek inneem van die oop porieë in die materiaal, en tweede is die feit dat water 'n lagie op die partikeloppervlakte vorm wat ondeurlaatbaar is vir keroseen-basis ferrovloeistof Dit lei tot die vermindering in koste van die ferro-hidrostatiese skeidingstegnologie wanneer nat materiaal in plaas van droë materiaal gebruik word. Soos alreeds genoem, partikels groter as 2.8 mm kan nat behandel word sonder enige negatiewe effekte op die skeiding. Vir -8+4 mm partikels is daar 'n ferro-vloeistofverlies van 0.6 tot 0.14 kg/ton voer vir' n 0-10% voginhoud. Die hoeveelheid ferro-vloeistof per ton voer materiaal wat verlore gaan wissel tussen 0.73 tot 0.56 kg vir -0.85+0.5 mm tot -12+8 mm partikelgroottes, respektiewelik. Die toename in ferro-vloeistofverlies by kleiner partikels is as gevolg van die toename in die oppervlakarea van kleinpartikels vir ferro-vloeistof adsorpsie. Daar is 'n toename in porositeit wat gepaard gaan met 'n toename in hoeveelheid in ferro-vloeistof wat verlore gaan as gevolg van probleme met die herwinning van ferro-vloeistof wat binne-in partikelporieë vasgevang IS. Die byvoeging van water by groter materiaal verlaag die hoeveelheid . ferro-vloeistofwat verlore gaan as gevolg van verminderde porositeit. Die modellering van die hoeveelheid ferro-vloeistof wat verlore gaan, as 'n funksie van die partikelgrootte en porositeit, sal help met die skatting van die hoeveelheid ferro-vloeistof benodig om 'n sekere hoeveelheid materiaal te behandel. Dit is gevind dat die kwaliteit van die ferro-vloeistof wat herwin word dieselfde is as die wat aanvanklik gebruik is vir die materiaal skeiding. Dit wys dat die skeidingseffektiwiteit beïnvloed nie die gebruik van gehersirkuleerde ferro-vloeistof nie.