Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Fine Coal Dense Medium Cyclone

No description

Murray Nolan

on 17 February 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Fine Coal Dense Medium Cyclone

Two operating sites: Tertre (1957) and Winterslag (1965)
Although results were encouraging, the separations were still not of modern efficiency and sharpness requirements
The Dutch State Mines or DSM design DMC is still widely used today on coarse coal FUEL RESEARCH INSTITUTE - LANDAU MINE A dense medium circuit was erected and commissioned during first quarter 1977.
Raw coal was deslimed in a cyclone and screen dewatered leaving a fraction 0.5 - 0.075mm. This was mixed with heavy medium and pumped to a separating cyclone 150mm in diameter. Magnetite was recovered in a series of magnetic separators. The clean coal was thickened in a cyclone and screen dewatered. MAGNETITE DATA •1980. Fourie, P. J. F., Van der Walt, P. J., Falcon, L. M. The beneficiation of fine coal by dense medium cyclone. Journal of the South African Institute of Mining and Metallurgy.
•1987. King, R. P., Juckes, A. H. Performance of a Dense Medium Cyclone when Beneficiating Fine Coal. Coal Preparation, 1988.
•1993. King, C. J. The Fine Coal Heavy Media Experience at Curragh Queensland Mining Limited. Australian Coal Miner.
•1994. He, Y. B., Laskowski, J. S. Separation of Fine Particles in Dense Medium Cyclone: The Effect of the Medium Yield Stress. 12th International Coal Preparation Congress.
•1994. Kempnich, R. J., van Barneveld, S., Lusan, A. Dense Medium Cyclones on Fine Coal, The Australian Experience. 12th International Coal Preparation Congress.
•2002. De Korte, G. J. Dense Medium Beneficiation of Fine Coal Revisited. Journal of the South African Institute of Mining and Metallurgy.
•2007. Hart, G., Firth, B., Graham, J., Purdon, P. Screening and Magnetic Separations for Fine Coal Dense Medium Circuits. Australian Coal Association Research Program.
•2008. De Korte, G. J. Leeuwpan. Fine Coal Dense Medium Cyclone Plant. COALTECH. WITBANK COAL SEAM WASHABILITY DATA FINE COAL DMC FLOWSHEET
S.A. PILOT PLANT PLANT PERFORMANCES Fairly coarse magnetite used
Early separations promising, but magnetite losses of 2-3kg per ton of coal processed
Cut point a little difficult to control Finer magnetite used
New magnetics recovery system used after first test, capturing very fine magnetite
Vastly improved results in tests 2 and 3 indicate that fineness of the medium is of vital importance to the separation efficiency
Optimum medium size gradings found through extensive testwork and medium size variations PERFORMANCE - CONTINUOUS OPERATION Plant run continuously for ~70h

Feedrate: 5 t/h
Pulp Feedrate: 35,000 L/h
Operating Pressure: 85-150kPa
Cyclone Diameter: 150 mm CONCLUSIONS AND FINDINGS For sharp separations magetite must be at least 50% passing 10 microns, at this size a product containing 7% ash at up to 64% yields were possible with magnetite losses of about 1kg per tonne of coal fines treated
However if magnetite is too fine losses may become excessive (more modern magnetic separation circuits may help reduce this)
Due to the success of the trials, a fine coal DMC circuit was constucted at Greenside colliery for use on the same coal seam at similar feed rates and size distributions 1957 - Tertre, Belgium.
1965 - Winterslag, Belgium (17 years).
1983 - Homer City & Marrowbone Ck, United States.
1977 - Landau, South Africa.
1980 - Greenside, South Africa (18 years).
1992 - Curragh, Queensland.
1994 - Gregory, Queensland.
2008 - Leeuwpan, South Africa. FOLLOW-UP WORK - KING & JUCKES 1987 Described the parameters that characterise the partition curve in a dense medium cyclone: the imperfection, the cutpoint shift and the short circuits to both over and underflow The above figure is a schematic representation of the partition function for fine coal beneficiation in a DMC. The four operating characteristics are shown: cutpoint shift, separation efficiency (EPM) and the two short circuit errors. PARTITION FUNCTIONS Typical Partition curves for Greenside and Landau Mines. Figure 1 shows the variation of partition curve with particle size at a single medium density, figure 2 the normalised curves. Figure 3 illustrates the variation of partition function with medium density for a single particle size and figure 4 shows the effectiveness of normalising the curves HOMER CITY - 1983 Based on the pilot plant FCDMC work of Killmeyer in 1982, a full scale fine coal heavy media plant was constructed at Homer City CPP, Pennsylvania. This plant operated fairly successfully for several years using cyclones ranging from 150 to 650mm. During this time several conclusions were made:
While small cyclones give better performance at higher specific gravities (1.4-1.5) under pilot plant conditions, this was not necessarily true at lower specific gravities under plant operating conditions.
The media circuit requires adequate bleeds to remove contamination
Precise specific gravity control is needed for effective separation
A method of recovering and recirculating ultra fine magnetite is essential
Precise control over media to coal ratio is required
A sufficiently fine grade of magnetite is required

Not all of these features were incorporated and despite extensive modifications and performance reviews the FC DMC circuit was subsequently decommissioned after 5 years. Sadly the circuit at Homer City was never made to operate at the required levels of availability and consistency. Visits to Homer City as well as several other plants operating a fine coal DMC circuit by Curragh personnel convinced them that the operation of FCDMC circuits were no more difficult than normal heavy media ones and the questions concerning operating parameters were resolved. And so... CURRAGH - 1993 Curragh personnel, now armed with knowledge from their visits to Homer City set about designing their own FCDMC circuit with two critical factors in mind:
Slimes in the DMC feed increased the slurry viscosity which in turn hampered the separation
Ultra-fine magnetite was essential for good operation but also easily lost from the circuit in various ways
So incorporated in the circuit design were:
Complete desliming of the DMC feed
Highly advanced and thorough magnetite recovery systems

CLI corporation who had designed and operated Homer City washery, along with Sedgman and associates were chosen for the conceptual design of Curragh's new FCDMC circuit. The plant design included several key features:
Ability to make up fresh magnetite solution to the fines HM circuit by removing the finest magnetite from the coarse HM circuit and 'swapping' it with the coarsest magnetite from the fines HM circuit
A blackwater bleed from the coarse HM circuit into the new fines circuit to maintain a healthy proportion of sub-micron slimes to aid stability of the dense medium
Complete desliming of the fine coal feed to achieve desired separation and keep magnetite consumption to reasonable levels
Ability to bleed coarse media from the operating circuit for cleaning. This removes non-magnetic slimes which would otherwise build up and result in increased viscosity and poorer separation. CURRAGH CURRAGH FC DMC DESIGN The circuit chosen contained: CURRAGH COAL WASHABILITY Sampling around the flotation circuit during 1986 showed that a thermal coal fraction could be recovered from the tailings. In 1987 the circuit was modified so that a coking coal was produced from the first three cells and a steaming coal product from the last two.
In was eventually concluded that installation of a heavy media cyclone plant to treat the 0.5-0.125mm fraction and use of flotation to treat the 0.125-0mm fraction would be more economical and the overall treatment more efficient than the installation of additional flotation capacity to treat the 0.075-0mm material.
This can be evidenced in the next slide which shows the distribution of material with respect to specific density in the plant feed (0.5-0.125mm). Also shown is the cumulative ash versus specific density on the right hand y-axis. It can be seen that the coal was suited to making a low ash product with fairly easy separation. HMC were chosen over spirals as they allowed higher overall plant yields by recovering more from the coarse circuits, where considerably greater amounts of near gravity material were present at the separation density (than flotation circuit). A feed cleaning circuit designed to process the 0.5-0mm fraction of the plant feed to remove as much of the sub 0.125mm material as possible. This consisted of 500mm diameter classifying cyclones followed by sieve bends/horizontal screens to remove the slimes remaining in the cyclone underflow.
An operating loop in which feed was mixed with correct dense media in a wing tank sump and pumped to two 500mm diameter conventional 'DSM' style heavy media cyclones at 50kPa.
A correct media recovery stage where the overflow and underflow from the heavy media cyclone was processed on vibrating sieve bends to remove most of the correct dense media which was returned to the correct dense circuit.
A product recovery stage where the overflow from the vibrating sieve bends mentioned in the correct media recovery stage was repulped by the addition of blackwater and sent to specially designed magnetic separators to remove most of the remaining magnetite. The separator effluent was then pumped to 600mm diameter thickening cyclones and the underflow retreated by a second stage of magnetic separators before final dewatering. The clarified water was all recycled.
A correct media loop which received the vibrating sieve bend underflows as well as the thickened magnetite from the magnetic separators. The density of the media was controlled by water injection to the pump inlet feeding distributor box providing make up to the operating loop. A nuclear density gauge on the feed line to the box monitored the density.
Rejects were dewatered by passing them over a conventional sieve bend and horizontal vibrating screen before disposal with the coarse reject. Product was mixed with the coarser fraction of flotation product and dewatered by screenbowl centrifuge. CURRAGH FCDMC RESULTS CONCLUSIONS & TROUBLESHOOTING CONCLUSIONS The fine coal DMC plant at Curragh achieved an ash content of 7% running on a coal feed of 0.5-0.125mm with better than expected magnetite consumptions.
Problems however were experienced in providing a feed free of -0.125mm material and so separating efficiencies were not quite as good as expected, with improvements expected to meet or exceed these planned at the time of the 1993 report. The feed pressure of 45kPa (a traditional 9*diameter) was also found to be too low. PROBLEMS and SOLUTIONS Aside from the inability to remove all the sub 0.125mm material from the feed, the fine coal DMC plant at Curragh also experienced large amounts of very fine pyrite in the feed, which passed through the classifying
cyclones, stuck with the product coal and were not removed by mag seps and so reported to the final product.
These problems were dealt with by increasing the size of the sieve bend apertures and providing a bleed off from the magnetic separator circuits so the pyrite could be sent to flotation, where it would correctly end up as reject material.
Additionally the DMC pressures were increased to 120kPa and the apex size increased, although at the time of the report this work was still ongoing so as to find the optimum characteristics.
Wear on the sieve bends was greater than expected and a continuing source of circuit inefficiency. It was difficult to avoid without to entirely replacing the screen deck. Equations, Equations, Equations... AN ASIDE... As the size of a DMC and thus the dense medium itself decreases, the separation efficiency of cyclones becomes heavily affected by the viscous drag on the fine particles. As magnetite concentrations increase the media can begin to behave as a non-newtonian, yield-thinning fluid, and here the yield stress greatly influences the dense medium separation.
To verify this He and Laskowski ran extensive tests on a 150mm DMC loop using fine density tracers. The medium properties were varied, using different grades of magnetite and a wide density range. They found that the medium stability, separation efficiency and cutpoint shift are all closely interrelated with the yield stress of the magnetite suspensions.

When magnetite suspensions become unstable, the magnetite itself succumbs to the centifugal forces in a cyclone and segregate, forming density gradients within the cyclone. This in turn, causes the magnetite to segregate, with most reporting to the underflow, a phenomena measurable by comparison between underflow and overflow density differential.

Bingham equation....Rheological properties....viscous friction....Reynolds number....Casson equation....Non-newtonian....Stoke's law....shear rate....Plastic liquid....yield stress....etc... A RESULT OR TWO... ...OR THREE... ...OR FOUR... CONCLUSIONS Medium stability, as characterised by the density differential in a DMC, is closely interelated with the Casson yield stress; it decreases exponentially with increasing Casson yield stress.
The separation efficiency of fine particles is also closely correlated with the Casson yield stress. Ep values show a continuous increase with Casson yield stress; the increase becomes more significant for finer feed particles -ie. below 0.5mm. The effect of medium rheology on separation efficiency of coarse particles -ie. 2 to 4mm is not significant.
The cutpoint shift for fine particles -sub 0.5mm is directly affected by viscous drag; increasing the Casson yield stress of the medium, which is associated with a higher viscous drag on separated particles, results in a greater cutpoint shift. The cutpoint shift for coarse feed particles -above 2mm is more sensitive to the influence of medium stability; increasing yield stress improves medium stability and reduces the cutpoint shift.
The results showed that yield stress played a very important role even in the dynamic DMC separation process. Particle movement in non-Newtonian, yield-thinning fluid is a function of both viscosity and yield stress; the movement of near-density fine particles conforms to Stoke's law and is mainly determined by the yield stress. FYI CASSON YIELD STRESS (CASSON MODEL OR CASSON EQUATION): A commonly used rheological model that quantifies yield stress and high shear viscosity, typically used for inks and molten chocolate. BINGHAM MODEL (OR BINGHAM EQUATION): A simple rheological model that relates shear stress and shear rate and quantifies yield stress and high-shear viscosity. NON-NEWTONIAN FLUID: A fluid which exhibits a viscosity that is dependent upon the shear conditions. SHEAR THINNING: Viscosity decrease with increasing shear rate. Shear-thinning is the most common form of non-Newtonian behaviour and is seen in suspensions, emulsions, polymer solutions and gels. The change in viscosity with shear rate for some materials can span several orders of magnitude, thus emphasising the adage that for most products “viscosity is a plot, not a dot”. YIELD STRESS: A structured fluid, such as an emulsion or suspension, possesses a yield stress. This is the stress that must be applied to disrupt internal structure (due to colloidal or other interactions) and elicit a significant decrease in viscosity. The presence of a yield stress can imbue products with desirable handling, appearance and storage properties. Yield stress can be measured using both oscillatory and viscometric sweep methods. KEY TO CURRAGH'S SUCCESS CLIMAXX Counter-Rotational Drum While similar to a conventional counter-rotational drum, the CLIMAXX featured:
750 Gauss magnetic field strength
an additional row of magnets, giving 11 interpoles total
lower than normal drum speed of 8-10 rpm
new tank design incorporating feed distributor and full width effluent overflow.
Excellent magnetite recoveries were achieved by these drums, even in the presence of high concentrations of non-magnetics in the feed. Concentrate densities in the range of 2.1 to 2.2 were consistently achieved. Its is important to note in Figure 4 above, that magnetite consumption showed no marked increase post FC DMC circuit commissioning. EFFICIENCY IMPROVEMENTS CURRAGH FOLLOW-UP - 1994 Curragh CHPP flowsheet A step by step approach was taken to alter parameters and improve DMC performance.
Attention was paid to controlling the magnetite size distribution. Despite extremely fine magnetite grades (35% passing 7 microns) density differentials were controlled to less than 0.3RD. However this did not reduce the product ash which was initially too high, or significantly improve the partition curve.
Cyclone geometry issues were also addressed. Several spigot sizes were trialled but to little success.

Eventually it became evident there was insufficient energy to effect a sharp separation so inlet pressure was increased to 65kPa and this showed improvements on the +0.25mm fraction but not the 0.25-0.125mm fraction. The units were then converted to a single 500mm cyclone and the inlet pressure increased to 120kPa. New spigot sizes were trialled and optimised at 195mm. Further increases in pressure with 'normal' spigots were tested but showed no further improvements.

Due to the marked improvements in efficiency at 120kPa these conditions were maintained and established as the basis for operation. The target ash of 7% was achieved consistently from this point. NOTE: High misplacement of reject to fines CURRAGH FOLLOW-UP CONCLUSIONS The Curragh project was not a straightforward application of existing technologies rather it was a path of progressive changes and learnings to ultimately produce a successfully operating DM circuit for the beneficiation of fine coal.
"To that end, a foundation has been laid for the Australian application of new technology which will be of benefit to the industry as a whole"

Some technical conclusions drawn from the project include:
Providing that magnetite sizes distribution is maintained fine enough, Ep's were improved dramatically with increased pressure, due to the availability of sufficient energy to separate the fine coal.
The process design successfully maintained magnetite consumptions at very low levels. Overall magnetite consumption in the Curragh plant remained consistent with that prior to commissioning of the fine coal plant.
Low Ep's of 0.04-0.05 were achievable on the +0.25mm material. However significantly higher Ep's of 0.1-0.11 were achieved on the -0.25mm material. Overall Ep was measured to be 0.8 on average.

Furthermore it was hypothesised that by directing the -0.25mm material to flotation and increasing the topsize of the FCDMC feed to 1mm an Ep improvement can be made on both included fractions.

"In summary, cleaning of fine coal by DMC has now become a proven process and as such, can be incorporated into the industry where requirements dictate, and done so very successfully providing care and attention is paid to detail in the design" -Kempnich, van Barneveld & Lusan, 1994. SOUTH AFRICA REVISITED Motivation - 2002 By 2002 all FCDMC circuits in SA had been decommissioned with spirals being favoured for fines processing. However, Witbank No.4 seam could not be beneficiated via spirals with any degree of success and so a method of processing this seam with sufficiently high quality to allow the fine coal to be added to export product was needed.
Dense medium had the potential to beneficiate the fine coal from Witbank No.4 to above the desired quality and much more efficiently than spirals with better control over product quality.
DMCs had been used at Greenside S.A. previously for 18 years. They were capable of producing a low ash coal of 7% ash and a middlings fraction of 16% ash. Greenside had employed 150mm cyclones at 150kPa. Magnetite was 50% passing 10microns and although the process was difficult to operate it nevertheless produced the required product.
The need to process the Witbank No.4 fines efficiently led to a re-evaluation of FCDMCs as it was the only process available to provide the required separation. The process needed to be made more practical though, with larger cyclones, lower pressures and utilising commercially available magnetite while still offering sharp separations. THE DESIGN It was found through simulation studies that the Ep could be lowered significantly by employing three stages of cyclones rather than a single cyclone as had been used in all previous FCDMC plants. Thus, the design featured cyclones used in a rougher-cleaner-scavenger arrangement. The results of the original simulation work can be seen below. The chart shows that the overall Ep is greatly reduced when compared to the individual rougher, cleaner and scavenger units. The simulations also showed that the amount of recirculating coal could become a problem if it wasn't carefully controlled. As always, complete desliming at 100microns was necessary, as was the capture of magnetite.

The overflow from the rougher is mixed with circulating heavy medium and became the feed for the cleaner (floats) cyclone. The underflow from the rougher became the feed for the scavenger (sinks) cyclone. The sinks overflow and the floats underflow are recirculated back to the rougher feed. The alternate streams are passed through a magnetic separator, dewatered and form the reject stream. PLANT PERFORMANCE DATA 2002, To-date results These two initial tests showed that the plant was capable of producing coal of the required quality, although they were performed shortly after commissioning. The magnetite consumption was 1.5kg/tonne. DESIGN ISSUES ADDRESSED - 2007 A 2007 report written by CSIRO and ACARP (including our own Peter Purdon) stated that the problems with FCDMC circuits is not so much in the operating of the cyclones themselves, but in the preparation of the feed and capture of the media.
Developments in more recent years such as high intensity magnets have shown that magnetite losses can be kept absolutely minimal even without magnetic separator units in series and screen bowl centrifuges have shown excellent dewatering capabilities as well as rapid size classification in the 0.1-0.25mm size range.

The conclusions of this report were that the use of high intensity magnetic separators and centrifuges for classification can significantly improve the issues which have been identified as the major problems with a dense medium circuit for fine coal. The use of scroll centrifuges may have advantages or disadvantages when compared to more traditional technologies.
Overall, the use of FCDMC technology in washeries is an economic descision. Any increase in operating costs and difficulties will have to be outweighed by an increased revenue from the higher quality of the final coal product. LEEUWPAN - SOUTH AFRICA 2008 Based on studies showing DMC to be more efficient than spirals, the decision was made to install a FC DMC circuit at Leeuwpan. The plant consisted of two identical circuits each processing 30 tph of feed. A typical example of the initial results are shown in the table below.

The ash results in all cases are far too high, but these reflect the very high yields obtained and the reject ash values of up to 92%. These points aside, it was clear the plant was providing a very good separation, just at too high a cut-point, which could be controlled. Further investigations revealed a faulty feed control gate which allowed some raw coal to by-pass directly to the product and that the spigots on the cyclones were of standard size which caused high cut-point densities. The addition of larger spigots helped to reduce these cut-points. EFFICIENCY TEST - OCTOBER 2007 The results obtained indicate that the cut-point density was extremely high, despite low cyclone feed densities. The high yield and product ash support this. The high amount of misplaced sink in the floats seems to support the idea that the spigot could not handle the sinks. The chart on the next slide demonstrates this. CYCLONE SPIGOT TESTS In order to drop the cut-point densities the spigots on the cyclones were increased from 120 to 145mm for the primary cyclone and from 100 to 120mm for the secondary. In a series of tests, the circulating medium density was reduced progressively in order to test the new spigots. The results show that the circulating medium density needs to be lowered to 1.2RD in order to achieve the desired product ash value of 13%.

These tests also showed that ash content is directly related to the density of the circulating medium. The next slide demonstrates this. It is important to note that even relatively small changes in the circulating medium density cause large increases in the product ash percentage. EFFICIENCY TEST No.2 On the 14th of February 2008 a second efficiency test was performed. The same sampling procedure as previously was employed. The raw coal being used while testing occurred was of slightly better quality. Results follow: The results obtained once again indicated a high cut-point density. This time however, this was largely due to the relatively high circulating media density. Product ash of 16% corresponds to the product ash obtained at the same circulating media density as the previous set of tests. The Ep and organic efficiency values obtained were quite acceptable and amount of misplaced material also. MAGNETIC SEPARATOR EFFICIENCY The above results indicate that the magnetic separators were working very well, with an average consumption of approximately 1.3kg/tonne of feed. To the time of the writing of the report, the magnetite consumption had not increased and had remained within acceptable ranges. CIRCULATING MEDIUM The samples of circulating medium taken indicate that its was very coarse. This can be seen below. Of positive note is that the circulating non-magnetics were very low, indicating the magnetic separators in the discard streams were working very efficiently. During follow up studies of the circuit it was noticed that the ultra fine magnetite was lost from the system gradually and a magnetite 'switching' method was being installed (similar to the one mentioned at Curragh). Since it is known that magnetite size is critical to successful separation, this was assumed to allow easier control of efficient operation of the FC DMCs.
Remembering from previously that 50% passing 10microns was considered close to ideal for efficient separations in FCDMCs, the 53 and 76% passing -45 seems far from adequate. FEED SIZING The above results indicate that the feed size is highly variable. The amount of sub 100 micron material is also quite high. Particle size distribution is important to be aware of as it affects both the cut-point density and Ep. Finer coal results in a higher cut point densities and thus higher Ep values. Any included sub 100 micron material would not be beneficiated effectively and so would raise the ash content of the product and hence, is removed from the feed with deslime screens prior to entering the cyclones. A lot of work was being done on screening methods to completely remove all the 100 micron material at the time of writing of the report. CONCLUSIONS The FC DMC plant at Leeuwpan was operated continually without issue and measured efficiency was within expectations. The plant was capable of producing the required product quality. Further work was being undertaken on improving and controlling cut-point density, magnetite size consist and desliming of the feed. SOME OVERALL CONCLUSIONS FINAL CONCLUSIONS DRAWN In the opinion of the presenter, fine coal beneficiation by dense medium cyclones is possible across almost all types of coal seam, producing quality products with as little as 7% ash or better. Dense medium can also be used to process coals that otherwise could not be handled by spirals. However, the process is inherently more difficult to operate than spirals and its use should be first checked by a cost benefit analysis. FC DMCs work extremely well, but require very close monitoring and strict control of operating characteristics such as:
Feed desliming - recommend a total removal of all sub 100 micron material, potentially to flotation.
Circulating medium size and concentration - medium should be at least 50% passing 10 microns and at an approximate medium to coal ratio of 3:1
Inlet pressure - Related closely to spigot size, but should generally be above the 9*diameter normally associated with coarse coal DMCs. 120kPa appears to work well.
Spigot size - should be larger than what is considered 'normal', but should be sized appropriately according to feed pressure in order to keep cut-point down.
Cut-point - needs to be closely controlled. Can be varied via spigot size, inlet pressure, medium size and density .
Magnetic separation - Needs to be able to recover almost all magnetite to keep running costs down and ultra fine magnetite in circulation. This also means that contamination of the recirculating media can be kept to a minimum.
G-forces - Should be kept between 200 and 300 as performance has been shown to drop off below and above these values, respectively. THANK YOU. ANY FINAL QUESTIONS?
Full transcript