Acid grade fluorspar which is in great demand by the chemical and aluminum industries, must contain at least 97.5% CaF2 with not more than 1.5% SiO2 and 0.5% Fe2O3. Often the Silica is limited to 1.2% with penalties starting at 1.0% SiO2. These limitations on grade and impurities require extremely close mill control, particularly through flotation where selectivity and high recovery is essential.
Over 95% of all acid grade fluorspar is processed by flotation through Sub-A (Fluorspar Type) Flotation machines. These machines, of the cell to cell type, are designed special for fluorspar with a high degree of flexibility essential for selectivity and multiple cleaning of concentrate. Middlings and clean tailings often must be completely isolated from the separate cleaning steps and diverted to the proper part in the milling circuit for most economical and efficient retreatment.
The flowsheet illustrated above is typical for the average Sub-A Fluorspar Flotation mill treating up to 100 tons of mine run ore per 24 hour day. Actual flotation conditions and equipment requirements should always be determined by having a comprehensive test made on the ore before proceeding with any fluorspar operation. Fluorspar ores may be quite complex, particularly when associated with lead and zinc sulphides, barite, calcite, iron oxide, and siliceous impurities. For this reason, a laboratory flotation test should be the first step in establishing a flowsheet.
For the average small mill treating up to 100 tons of ore a day, primary crushing is usually adequate and very economical. Larger tonnage will require primary and secondary crushing for maximum efficiency in size reduction and subsequent ball milling.
Fluorspar ores usually require grinding to 48 or 65 mesh to liberate the calcium fluoride from the gangue impurities. Ball mill grinding with a Steel Head Ball Mill in closed circuit with classifier is the general practice. In larger plants, particularly when fine grinding is necessary, thickening of the classifier overflow is necessary to maintain proper density and feed regulation to flotation. This thickening step on fluorspar ores containing sulphides is usually between the sulphide and fluorspar flotation circuits. Reagents used for selective flotation of lead and zinc then can be rejected in the thickener overflow water.
Normally, conditioning at mill temperature willthoroughly film the fluorspar with reagent and makeit readily amenable to separation and recovery by flotation. Heating the pulp, even up to the boiling point,is often advantageous.
AnAgitator and Conditioner is ideal for fluorspar conditioning as circulation is positive and thorough reagentizing with a minimum amount of reagent is assured. Any frothing tendency is dissipated in the pulp through the stand pipe and adjustable froth collar.
Flotation of fluorspar must be extremely selective when producing acid grade concentrate. This selectivity is essential as the ratio of concentration is low, often up to 80% or more of the entire tonnage, and must be floated in the rough circuit. Cleaning by two or more stages of flotation must bring the rougher product up to acid grade and at the same time retain a high weight recovery with a minimum circulating load.
The Sub-A Flotation machine, the accepted standard in all fluorspar flotation plants, has been adapted specially for fluorspar treatment with provision for multi-stage cleaning and recirculation of middling products without the need of auxiliary pumps. Cleaner tailings may be conveniently removed at any point in the circuit. The flowsheet on the reverse side of this page shows one of the many possible cell arrangements used in treating fluorspar ore.
Thickening of fluorspar concentrates offers no special problem. Thickener capacity, however, should be adequate to handle the tonnage and have ample storage capacity during possible interruption in the filtering and drying sections. Fluorspar flotation froth has a tendency to build up on the thickener surface, but this can be taken care of by retaining rings near the overflow lip and by sprays so only clear water overflows the thickener. Thickened concentrates at 50 to 60% solids is removed by a Adjustable Stroke Diaphragm Pump, feeding by gravity to the filter.
Fluorspar is extremely rapid filtering even when ground fine, provided a non-blinding filter media is used. The rotary fluorspar type filter with stainless steel filter media, heavy duty oscillating mechanism, oversize valve and ports, and high displacement vacuum pump is standard for fluorspar flotation concentrates and will discharge a filter cake with as low as 6% moisture. In the event the filtrate is slightly turbid or contains solids, it should be diverted back to the thickener. For this reason a adjustable stroke diaphragm pump is often used in place of the conventional centrifugal filtrate pump.
Fluorspar flotation concentrates of acid grade must be dried to less than 0.5% moisture. Dust losses are kept to a minimum by providing a closed system with a cyclone to insure only vapor laden air discharging to the atmosphere. Enclosed screw conveyors, elevators and often air-born systems are used to transport the finely divided dried acid spar to the storage bins. Provisions should be made for handling efficiently the hot concentrate discharging from the dryer. The Standard Dryer is ideal for this purpose.
Fluorspar ores often contain appreciable amounts of sulphides in the form of galena, sphalerite, or both. These sulphides, when present, not only represent a valuable constituent of the ore, but also must be removed prior to fluorspar flotation to meet the market specifications for acid grade fluorspar.
If lead and zinc were present, the same flowsheet would apply to remove a bulk sulphide concentrate which could be subsequently refloated to produce the respective lead and zinc concentrates suitable for marketing.
The best approach to effectively produce separate lead and zinc concentrates should be established by test work. In some cases, selective flotation is indicated initially. This may be accomplished by removing a lead concentrate, then following this process by conditioning and flotation of the lead tailing to produce a zinc concentrate.
Conditioning of the classifier overflow is required if sulphidization is employed to effect flotation of oxidized lead. A second stage conditioning of the thickened lead tailing, after repulping with fresh water, is required for flotation of the fluorspar. Heating of the pulp at this point is often advantageous.
The lead and fluorspar are recovered by Flotation of the cell-to-cell type, permitting maximum recovery and grade of concentrate. Wide acceptance of machines is well verified when considering that over 95% of all acid grade fluorspar is processed in the Sub-A Flotation Machine. Flexibility of these machines is of prime importance where such high specifications must be met. Multiple cleaning, always necessary in acid grade fluorspar plants, can be performed without the help of pumps.
Both concentrates are thickened and filtered. The thickenedlead concentrate is filtered on the Disc Filter. Thickened fluorspar concentrate, at approximately 60% solids here, has a high filter capacity of approximately 2000 pounds per sq. ft. per 24 hours. The Fluorspar Filter with its stainless steel filter media, is especially designed for this application.
The Standard Dryer effectively dries the filtered fluorspar concentrate to less than 0.5% moisture, as required for marketing. An elevated temperature in the dryer can also be used to burn off small amounts of sulphur and lead.
A screw conveyor and bucket elevator as employed to transport the dried fluorspar to the concentrate storage bins. Bins can be conveniently discharged into rail road cars for shipment, while the filtered lead concentrate may be marketed as produced, without drying.
While many ores respond to the same general pattern of treatment, each ore is an individual problem.Such is the case of this fluorspar ore which is characterized by the presence of a portion of the fluorite inextremely close association with calcium carbonate andsilica and containing appreciable clay.
High acid grade fluorspar concentrates are difficult to obtain from this class of ores by flotation with an ordinary -65 mesh grind. The concentrates, in this study, are currently being used for production of hydrofluoric acid and synthetic cryolite. Market requirements demand that the calcium carbonate content be reduced to an absolute minimum. Moreover, the future productionnow in demand, is desirable. This study deals with a flowsheet designed to achieve high recovery of acid-grade fluorspar in an economical manner.
The typical fluorspar flotation flowsheet normally consists of stage grinding by ball mill in closed circuit with a mechanical classifier followed by conditioning of the pulp either with or without steam in the presence of reagents followed by Sub-A Flotation with three or more cleaning steps by reflotation. This particular ore does not, with the normal flowsheet, produce an acid grade concentrate of 97.5% CaF2 with less than 1.5% SiO2.
The ore being studied is crushed underground at the mine and partially beneficiated by the heavy media process. This washed ore is further crushed at the mill. Soda ash is added to the primary grinding mill which is in open circuit with a duplex Spiral Classifier. The classifier is in closed circuit with the secondary grinding mill and the classifier overflow, which is all 65 mesh, is pumped by a SRL Pump to the Conditioner where the following reagents are added:
Reagent Amount, Pounds per ton Na2Si03 (Optional) 0.2 Soda Ash 2.0 Oleic Acid up to 2.0 Quebracho 0.2
The conditions presented by this particular ore illustrate the importance of complete laboratory investigations as a great many different combinations of treatment were required to develop the final flowsheet. The deviations from the standard fluorspar flowsheet were first substantiated by locked cycle batch laboratory tests followed by a small tonnage pilot plant run to verify the laboratory results before final recommendations were made.
The rougher flotation circuit produces a final tailings while the rougher concentrate is subjected to the first cleaning stage. A 6 cell Sub A Flotation Machine, cell to cell type, is used for the rougher flotation and 6-cell Sub-A Flotation Machines are also used for the three cleaning steps.
Tailings from the first cleaners are pumped to a Morton 2-stage Cyclone for the removal of clay slimes. The ability to add clear water for washing in the classifier makes the Morton Cyclone particularly useful at this point in the flowsheet. The slimes go to final tailings and the cyclone sands, at high density, are reground in a Regrind Mill which is in closed circuit with a Hydro-Classifier. The regrind is to 325 mesh and the hydro-classifier overflow returns to the first cleaner cells for reflotation. Reagent sodium silicate is recommended to aid classification. Concentrates from the first cleaners go to the second cleaner cells where further up-grading takes place.
The middlings (tailings) from the second cleaner cells go to the hydro-classifier in the re-grind circuit. The concentrates from the second cleaners advance to the final cleaners. Tailings from the final cleaner cells are returned to the second cleaners and the final, high grade concentrates are filtered, dried and shipped to market.
The concentration of fluorspar ores for the production of acid grade concentrates is accomplished by the use of combinations of reagents such as pH regulators, depressant and fluorspar promoters. The reagents commonly used are as follows:
Factors of simplicity, initial low plant cost, together with flowsheet flexibility for maximum results on a difficult ore were basic considerations in the design of this 125-ton per day Fluorspar Flotation Mill. The design proved successful and accomplished the desired metallurgical results, with low capital expenditure and operating costs.
Following numerous laboratory tests, a flowsheet was developed that gives flexibility to handle the several types of fluorspar ores. Two stage open circuit crushing, with the average ore ground to 100 mesh,gives maximum results. Fine grained ores with some sulphides require secondary classification and a sulphide flotation stage. Due to character of most fluorspar ores heating the pulp gave improved results, and necessitated the installation of a boiler to provide hot dilution and make up flotation water for five stages of cleaning and recleaning. A Apron Feeder controls the feed from crude ore bin to jaw crusher while a wedge bar grizzly ahead of the jaw crusher removes the fines from the crusher feed. A 2x4 Dillon Screen removes the fines ahead of secondary crushing. An adjustable stroke ore feeder controls feed to the 5x8 Steel Head Ball Mill, and the spiral classifier discharge is pumped direct to flotation section or to hydroclassifier for secondary classification, depending on requirements.
The machinery was located for accessibility, ease of operation, minimum loss of floor space, resulting in reduced size of mill. The crude ore bin was constructed of natural timber on the site, on a steep slope, reducing expense of excavation and construction. An 8 clear opening rail grizzly prevented oversizegoing into bin.
The buildings for crushing section and mill are of light steel construction with corrugated sheet metal on walls and roof. The frame work and trusses lightweight for buildingsupport only and provided without insulation, because of mild climatic conditions. Account of heavy snowfalls the roof slopes are all of quarter pitch.
Launders on cleaning stages are made so that flows can be changed to regulate number of cells required, depending on the type ore being treated. Wood platforms and walkways of 2 spaced lumber are used in flotation sections, while piping between machines is carried below the floor.
All electric lighting and power wiring with ample reserve are in rigid conduit with flexible connections to motors; and motor controls are mounted on wall panels with stop and start push button stations located within sight or near each motor. Fluorescent lighting is provided over flotation section, as it gives operators better visual control of the flotation operation.The Rotary Dryer is lined with fire brick at discharge (burner) end.
With depletion of high-grade deposits, production must depend upon low-grade deposits that are highly contaminated with impurities which may be silica, calcite, barite, iron oxide, and sulphides such as pyrite, galena, and sphalerite, in close association. The flotation problem is largely one of impurity removal. The sulphide minerals are generally floatedfirst, and then the fluorite is floated from the silica, calcite, and other impurities.
Oleic acid or various mixtures of oleic and linoleic acids with soda ash and sodium silicate as silica depressant and slime controller, and quebracho to depress calcite, are the common reagents for fluorspar flotation. Sometimes pre-sulphide flotation with xanthate and a frother is necessary to remove sulphides and, often, heating the pulp to boiling temperature is advantageous in effectively depressing the silica, calcite and other associated minerals in the cleaning stages.
[/fusion_builder_column][fusion_builder_column type=1_1 background_position=left top background_color= border_size= border_color= border_style=solid spacing=yes background_image= background_repeat=no-repeat padding= margin_top=0px margin_bottom=0px class= id= animation_type= animation_speed=0.3 animation_direction=left hide_on_mobile=no center_content=no min_height=none]Geology-Where-are-Fluorspar-Deposits
This report is the fourth in a Bureau of Mines series describing the sodium fluoride-lignin sulfonate-fatty acid process of froth flotation separation of fluorspar from complex ores containing fluorspar, barite, calcite, and quartz which was developed and patented by Clemmer and Clemmons of the Bureau of Mines. At the Tucson (Ariz.) Metallurgy Research Laboratory the ores of Arizona were studied; and at the Tuscaloosa (Ala.) Metallurgy Research Center, the ores of Kentucky, Tennessee, and Illinois were studied.
The sodium fluoride-lignin sulfonate-fatty acid process is applicable to a variety of ores of different grades and mineral association for recovery of fluorspar from associated gangue materials; it has been shown to be practicable in continuous pilot plant operation as well as laboratory-scale flotation tests. This report deals with the application of the process to a complex calcareous fluorspar ore from Illinois and presents the results of laboratory batch flotation tests and continuous pilot plant flotation tests for recovery of the fluorspar in the ore.
The largest use of fluorspar is in the production of hydrofluoric acid in which no satisfactory substitute for acid-grade fluorspar is known. A prospective new outlet for hydrofluoric acid is in its addition to the oxidizer of the Atlas rocket, which will significantly increase the booster performance. The second major use of fluorspar is as a flux in the manufacture of basic open hearth and basic electric furnace steels in which no suitable materials are available to replace metallurgical-grade fluorspar, A third use of fluorspar is in the manufacture of glass and ceramic products. The specifications and prices of the various grades of fluorspar are listed in appendix A.
The complex fluorspar ore used in the investigation was from the fluorspar district near Cave-in-Rock, III. ; a 14-ton sample of ore was obtained from the Minerva Co. Crystal mine located about 5 miles west of Cave-in-Rock.
Petrographic examination showed that about 38 percent of the fluorspar reporting to the minus 48- plus 65-mesh fraction contained inclusions and that about 24 percent of the fluorspar was locked in the minus 325- plus 400-mesh fraction. However, the carbonate and quartz crystals locked in the fluorspar mineral were extremely small; in a concentrate analyzing 98.0 percent CaF2, 30 percent of the fluorspar grains were locked. The petrographic analysis revealed that no appreciable benefit to mineral liberation would be achieved by crushing finer than 65 mesh.
The primary carbonate in the ore was calcite with a considerable quantity of dolomite. The silica present was reported as quartz. Other materials consisted of 1.3 percent sphalerite and minor amounts of barite and galena. A chemical analysis of the sample is shown in table 1.
Samples of the ore were prepared for flotation by dry crushing to minus 10 mesh followed by wet stage grinding to minus 65 mesh in a laboratory pebble mill, using Tuscaloosa city tap water that had about 45 parts per million equivalent calcium carbonate total hardness. Prior to flotation, the ground ore pulp was treated, at about 40 percent solids, in a mechanically agitated flotation cell with conditioning reagents and then with a collector. A rougher fluorspar concentrate was floated off and cleaned six times.
A series of preliminary flotation tests was made of the ore to determine the quantities of sodium fluoride and calcium lignin sulfonate necessary to produce the maximum recovery and grade of fluorspar. The quantities of sodium fluoride and calcium lignin sulfonate were varied from 2.0 to 8.0 pounds per ton of ore; the quantity of oleic acid was held constant at 0.48 pound per ton of ore. The grade of fluorspar concentrate was increased with the dosages of sodium fluoride and calcium lignin sulfonate and leveled off at 5.0 pounds per ton. The summarized results of flotation tests made to determine the effect of varying the quantities of sodium fluoride and lignin sulfonate are given in table 2.
The laboratory batch flotation studies were continued to determine the optimum quantity of collector needed to obtain the maximum grade and recovery of fluorspar. The pulp was conditioned (1) with 5.0 pounds of sodium fluoride per ton of ore. to disperse the pulp and clean up the mineral faces, (2) with 5.0 pounds of calcium lignin sulfonate per ton of ore to coat the surfaces of the gangue particles and render them hydrophilic, and (3) with various quantities of sodium oleate, as a collectors to concentrate the fluorspar. In most instances an acid-grade fluorspar concentrate was obtained. The rougher concentrate contained 94.3 percent of the total fluorspar in the ore at a collector dosage of 0.30 pound per ton of ore, but the mineral particles did not adsorb enough collector to sustain their flotation during the six cleaning stages. About 0.50 pound of collector per ton of ore appeared to be the optimum dosage; the grade and recovery of fluorspar were essentially constant with larger quantities. This indicated that large quantities of sodium oleate were adsorbed by the fluorspar mineral and not by the gangue materials. The summarized results of these tests are shown in table 3.
Another series of tests was made using various quantities of oleic acid as the collector while maintaining the quantities of sodium fluoride and calcium lignin sulfonate at 5.0 pounds per ton of ore. The tests revealed that the oleic acid was as selective as the sodium oleate in producing acid-grade fluorspar concentrates ; however, the fluorspar recovery was somewhat lower with the oleic acid because it did not disperse, as well. The optimum amount of oleic acid was 0.48 pound per ton of ore. The summarized results of these tests are shown in table 4.
Additional laboratory batch flotation tests were made using the data obtained in determining the optimum quantities of reagent. The minus 65-mesh pulp was conditioned at about 40 percent solids in a mechanically agitated flotation cell for 5 minutes with 5.0 pounds each of sodium fluoride and calcium lignin sulfonate per ton of ore for dispersion of pulp and retardation of gangue minerals. Sodium oleate, 0.5 pound per ton of ore, was then added as a collector; conditioning was continued for another 5 minutes. The rougher concentrate was floated and refloated (cleaned) six times to remove gangue minerals. A concentrate analyzing 97.8 percent CaF2 and accounting for a fluorspar recovery of 84,5 percent was obtained. The results of a selected test are presented in tables 5 and 6.
Based on the results of the laboratory batch tests, a continuous pilot plant with a capacity of about 150 pounds of dry feed per hour was assembled. The process included grinding, classification, conditioning, and flotation, as shown by figure 1.
The ore was reduced by jaw and roll crushers in closed circuit to minus 3/8 inch and stored in a bin. From the bin it was transferred by a constant-weight feeder to a rod mill operated at 60 percent solids. The rod mill operated in closed circuit with a vibrating screen to grind the ore to minus 65 mesh. The screen undersize (minus 65-mesh) passed to a hydroseparator for
removal of colloidal slimes. The hydroseparator overflow represented about 1.5 percent of the weight of the ore and a loss of less than 1 percent of the total fluorspar. The hydroseparator underflow, at about 40 percent solids, passed to a conditioner where sodium fluoride and calcium lignin sulfonate were added. The discharge from the first conditioner flowed to a second conditioner where oleic acid was added as the fluorspar collector. A retention time of about 9 minutes in each conditioner gave satisfactory results. The conditioned pulp then flowed to a bank of three rougher flotation cells where a rougher concentrate was floated. The rougher tailing flowed to a single cell operating as a scavenger to recover additional fluorspar. The froth from this cell was recycled to the last rougher cell; the tails flowed to waste. The rougher concentrate was cleaned nine times, and the middlings were circulated back to the first cleaner where they were removed and thickened in a bank of three hydrocyclones (parallel). The underflow from the hydrocyclones was sent to the first conditioner, and the overflow went to waste. An emulsion-type collector (made up of 17.7 parts oleic acid 1.3 parts sodium oleate, and 361.0 parts water) was added to the second rougher cell to aid the flotation of the fluorspar.
The summarized results of a continuous flotation test are given in tables 7 and 8. The final fluorspar concentrate analyzed 96.4 percent CaF2 , a recovery of 90.0 percent of the total fluorspar in the ore. About 7 percent of the fluorspar was lost in the overflow from the cyclones.
The fluorspar concentrate was slightly below the specifications for acid-grade fluorspar; however, it meets all specifications for high-quality ceramic-grade fluorspar. It was possible to obtain an acid-grade fluorspar by introducing additional cleaners into the circuit; however, there was some sacrifice in recovery.
The laboratory batch and continuous pilot plant flotation tests demonstrated that the sodium fluoride-calcium lignin sulfonate-fatty acid method for selective flotation of fluorspar from complex calcareous fluorspar ore is an effective and practical means of producing high-grade fluorspar concentrates.
The flotation of the fluorspar in a continuous test in which the middlings were removed from the circuit, thickened, and returned for further conditioning produced a fluorspar concentrate analyzing 96.4 percent calcium fluoride, a recovery of 90.0 percent of the fluorspar in the ore. The fluorspar concentrate produced from a deposit near Cave-in Rock, III., meets all specifications for high-quality ceramic-grade fluorspar.
Developments in DAF plant design are summarized, especially those dealing with flocculation pretreatment and with DAF hydraulic loadings. Prior to about 1990, it was typical to design flocculation tanks much like flocculation ahead of sedimentation. Long flocculation tank detention times of 20 to 30 min with 2 to 3 stages of tapered mixing intensity were common. Using the principle obtained from the contact zone modeling presented in Section 3.3 that good performance can be achieved with small flocs of 10s of microns, Edzwald and co-workers demonstrated this through laboratory  and pilot-scale DAF studies . This led to the acceptance that floc tank design should be tailored to DAF plants, and that the flocculation tank effluent should produce flocs of 10s of microns, not flocs of 100s of microns as desired for settling processes. After 1990, flocculation times for many DAF plants have decreased to about 10 min, with at least one design at 5 min (Croton plant for New York City).
Again prior to about 1990, DAF tanks were sized for hydraulic loadings of 5 to 10 m h1. The considerable theoretical and applied research conducted on DAF in the last 25 years has led to a better understanding of the process. This knowledge and optimization of the process in design studies have pushed the technology so that much higher DAF hydraulic loadings are now used leading to very small footprint areas. In the 1990s, DAF plants were designed at 10 to 20 m h1. In the last 5 years, successful DAF operation at 20 to 40 m h1 with a short flocculation time of 5 min was demonstrated through pilot-scale studies [30, 31]. A DAF manufacturer has ingeniously developed a porous plate that is placed at the bottom of the DAF tank for better flow distribution in the separation zone and for collection of the subnatant . Several new plants have been built using this system, e.g., a facility in the USA (West Nyack, NY) designed at 30 m h1 and one in Tampere (Finland) at about 40 m h1. In certain situations where land is scarce or expensive, it is possible to reduce total water plant footprint area by placing DAF directly over the granular media filter. It does limit the hydraulic loading,, because the filtration controls the overall process. However, even with this design facilities have been built with a hydraulic loading of 15 m h1.
The case study in this paper considers a flotation plant that has eight species (Table1). The two-components model, named fast and slow, has been applied for chalcopyrite (CuFeS2), chalcocite (Cu2S) and pyrite. (FeS2). Table1 also gives the copper grade and feed mass flow rate.
For the superstructure, a flotation plant with six flotation stages was considered: rougher (R), two scavengers (S1 and S2), cleaner-scavenger (CS), cleaner (C1), and re-cleaner stages (C2). The superstructure is represented by the origin-destination matrix (see Table2). From Table2, it is possible to identify 1,200 circuit alternatives.
The mathematical model utilized is similar to the one used by Cisternas et al. (2014), this is a mass balance between flotation stages (bank), mixers and dividers, as it is shown in Figure1a. However, the bank model was based on a cells model (Figure1b) which allows to use a more complex model.
The design provided for the slurry produced by monitoring of the slimes dams to be pumped to a central treatment complex, comprising a flotation plant, uranium plant, three pyrite roasters, two acid plants and a gold plant. The pyrite concentrate produced by the flotation plant would be treated for the recovery of uranium by conventional acid leaching and then roasted for the production of sulphuric acid. Finally, gold would be recovered from the calcine.
Large quantities of water are required for the process, and the plant was therefore set up close to the S.A. Lands mine where such quantities were available from underground pumping. As the plant site was in an urban area, housing for the limited number of employees (900) was available locally and no mine townships were required. Full advantage was taken of the local infrastructure, but various problems were encountered in obtaining the required wayleaves for the extensive pipeline system that intersected railways, townships, motorways and other types of urban development.
The redeposition of the tailings required an area of some 750 ha, and a shallow valley surrounded by poor farming land sited approximately 11 km south of the plant was eventually selected. When all the dams currently scheduled for treatment have been removed, 1250 ha of prime land in urban areas will become available for development, while in addition a source of visual, air and water pollution will have been removed.
A period of 2 years was allowed for the design, construction and commissioning of the plant, the total cost being estimated at $165 million. Slime was first pumped to the plant in December 1977, and the operation formally came into production on 25 February, 1978, with the recovery of the first uranium. Sulphuric acid production started on 14 March, 1978, and the first bar of gold bullion was produced on 11 April, 1978, thus meeting the timetable laid down 2 years previously.
The total cost of the initial project, excluding stores inventories, was $166 million, less than 3% above the estimate made 18 months earlier. ERGO was developed during a slack period in the South African capital engineering industry, which meant there was no significant shortage of artisans, design and construction capacity or supplies of building materials and items of plant and equipment.
The simplest way of smoothing out grade fluctuations and of providing a smooth flow to the flotation plant is by interposing a large agitated storage tank (agitator) between the grinding section and the flotation plant:
Any minor variations in grade and tonnage are smoothed out by the agitator, from which material is pumped at a controlled rate to the flotation plant. The agitator can also be used as a conditioning tank, reagents being fed directly into it. It is essential to precondition the pulp sufficiently with the reagents (including sometimes air, Section 12.8) before feeding to the flotation banks, otherwise the first few cells in the bank act as an extension of the conditioning system, and poor recoveries result.
Provision must be made to accommodate any major changes in flowrate that may occur; for example, grinding mills may have to be shut down for maintenance. This is achieved by splitting the feed into parallel banks of cells (Figure 12.53). Major reductions in flowrate below the design target can then be accommodated by shutting off the feed to the required number of banks. The optimum number of banks required will depend on the ease of control of the particular circuit. More flexibility is built into the circuit by increasing the number of banks, but the problems of controlling large numbers of banks must be taken into account. The move to very large unit processes, such as grinding mills, flotation machines, etc., in order to reduce costs and facilitate automatic control, has reduced the need for many parallel banks.
Some theoretical considerations have been introduced (Section 12.11.2), but there is a practical aspect as well: if a small cell in a bank containing many such cells has to be shut down, then its effect on production and efficiency is not as large as that of shutting down a large cell in a bank consisting of only a few such cells.
Flexibility can include having extra cells in a bank. It is often suggested that the last cell in the bank normally should not be producing much overflow, thus representing reserve capacity for any increase in flowrate or grade of bank feed. This reserve capacity would have to be factored in when selecting the length of the bank (number of cells) and how to operate it, for example, trying to take advantage of recovery or mass pull profiling. If the ore grade decreases, it may be necessary to reduce the number of cells producing rougher concentrate, in order to feed the cleaners with the required grade of material. A method of adjusting the cell split on a bank is shown in Figure 12.54. If the bank shown has, say, 20 cells (an old-style plant), each successive four cells feeding a common launder, then by plugging outlet B, 12 cells produce rougher concentrate, the remainder producing scavenger concentrate (assuming a R-S-C type circuit). Similarly, by plugging outlet A, only eight cells produce rougher concentrate, and by leaving both outlets free, a 1010 cell split is produced. This approach is less attractive on the shorter modern banks. Older plants may also employ double launders, and by use of froth diverter trays cells can send concentrate to either launder, and hence direct concentrate to different parts of the flowsheet. An example is at the North Broken Hill concentrator (Watters and Sandy, 1983).
Rather than changing the number of cells, it may be possible to adjust air (or level) to compensate for changes in mass flowrate of floatable mineral to the bank. To maintain the bank profile at Brunswick Mine, total air to the bank was tied to incoming mass flowrate of floatable mineral so that changes would trigger changes in total air to the bank, while maintaining the air distribution profile (Cooper et al., 2004).
The main by-product source of uranium today is at Olympic Dam in South Australia, where low concentrations of uranium (0.025 to 0.050%U) occur with copper grading about 1.8%Cu. Present production there is about 3500 tonnes of uranium, but there are plans to increase annually this to 16 000 tU/yr.
Following primary crushing underground, the ore is ground and treated in a copper sulfide flotation plant. About 80% of the uranium minerals remain in the tailings from the flotation cells, from which they are recovered by acid leaching as in a normal uranium mill. The copper concentrate is also processed through an acid leach to recover much of the other 20% of the uranium. The pregnant liquor is then separated from the barren tailings and in the solvent extraction plant the uranium is removed using kerosene with an amine as a solvent. The solvent is then stripped, using an ammonium sulfate solution and injected gaseous ammonia. Ammonium diuranate is then precipitated from the loaded strip solution by raising the pH, and removed by centrifuge. In a furnace the diuranate is converted to uranium oxide product.
However, after the secondary recovery by acid leaching, some uranium remains in the copper concentrate as it proceeds to be smelted. Typically it would have 45% Cu and up to 0.15% uranium, and the uranium is recovered in the further copper processing. This creates a safeguards problem if the smelting and electro-refining is not done at the mine site.
In the past some uranium has been recovered as a by-product of phosphate production, and this is set to be revived, with new technology, on an increased scale. Phosphate rock (phosphorite) is a marine sedimentary rock, which contains 1840% P2O5 , as well as some uranium and all its decay products, often 70 to 200 ppmU, and sometimes up to 800ppm. The phosphate rock is treated with sulfuric acid to give gypsum and phosphoric acid, and the uranium is normally recovered from the phosphoric acid by some form of solvent extraction (SX). A new process PhosEnergy uses ion exchange (IX) and promises to reduce recovery costs significantly.
The potential amount of uranium able to be recovered from phosphoric acid streams is over 11 000 tonnes U per year (global P2O5 production in 2010 was 33.6 Mt). The economic benefit will be both in the value of the uranium and in reduced regulatory demands on disposal of low-level radioactive wastes arising from the process. Estimated uranium production costs will put the new process in the lowest quartile of new uranium production.
The process design using superstructure optimization has shown that there are cases in which the best structure is not highly sensitive to the operational values. For example, Cisternas and colleagues studied optimal structures for separation based on fractional crystallization (Cisternas, 1999) and found that, in many cases, the best structure was independent of changes in operating conditions. Furthermore, Cisternas and coworkers (Cisternas and Rudd, 1993; Cisternas et al., 2006) found that there are areas within a design region where a design is always superior to another, regardless of the operating conditions. Although this is not transferable to flotation circuit design, it sets a precedent for efforts to study whether this assumption is valid in the design of flotation circuits. Cisternas et al. (2004) developed a procedure for the design of flotation circuits based on mathematical programming using two-level hierarchical superstructures. The procedure allowed the flotation to be modeled using a first-order kinetic model. The method was applied to a copper flotation plant in several case studies; each one with different numbers of cells per bank and, therefore, different values of recovery in each flotation stage. The optimal flotation circuit obtained was the same, but with different values for the mass flow rates, overall recovery, and concentrate gradient. This result means that the same optimal structure results from different operational conditions and numbers of cells in each bank. Other cases were studied, including more complex structures, and similar results were obtained. Later, (Mndez et al. 2007, 2009) used different grinding circuits (including grinding, grinding-classification, classification-grinding, and classification-grinding-classification) in the design of flotation circuits. The application to a copper flotation plant indicates that the same flotation structure was obtained using different grinding circuits. These studies show that the optimal flotation circuit strongly depends on the feed composition and metal price but has a low dependence on the stage recovery. Jamett et al. (2012) presented a model for the design of flotation circuits with uncertainty using stochastic programming. Uncertainty was represented by several scenarios, including changes in the feed grade and metal price. The model allows for changing the operating conditions (residence time) and flow sheet structure for each scenario while maintaining fixed equipment design (number of cells in each bank of flotation) for all scenarios. The results showed that the optimal flow sheet structure did not change for 8 of the 9 scenarios studied, but the recoveries of each stage changed for each scenario. Additionally, Montenegro et al. (2009), Montenegro et al. (2010) and Montenegro et al. (2013a) studied the effect of the uncertainty in the recoveries of the rougher, cleaner, re-cleaner and scavenger stages on the global recoveries and final concentrate grade, among other indicators, for 12 flotation circuits. The uncertainties in the recoveries of each stage were represented by normal, triangular and uniform distribution functions with variations between 1% and 10%. In other words, the recoveries at each stage were not modeled with any kinetic model but were represented by distribution functions. The uncertainties were studied by considering the variation in each stage as well as in several stages simultaneously; 84 cases were considered. Monte Carlo simulation was used in the study, running more than 6 million simulations. The results showed that the best flotation circuits were not a function of the stage recoveries, i.e., for different values of stage recoveries, there is a set of flotation circuits that performs the best. Later, Montenegro et al. (2013b) applied a shortcut computational method to analyze and compare alternative flotation circuits to treat high-arsenic copper ores. Twenty-seven circuits were evaluated based on the metric indices of efficiency, capacity, quality, economic and environmental impact. The simulations were performed for an Australian sulfide ore containing chalcopyrite, tennantite, quartz, and pyrite. In the simulation, a constant stage recovery was assumed. To validate this assumption, the normalized indicators were calculated for several values of stage recoveries for each mineral; three levels were selected at 5, 10, and 20%, which can be considered moderate, intermediate and high variation in the stage recoveries. A random sampling of the case studies was selected to reduce the sampling error. The size of the sample was estimated using 28 combinations, which gives a 0.95 confidence level. Twenty-eight combinations were studied for twenty-seven circuits; therefore, 2,268 simulations were performed. The results of the simulations for moderate, intermediate and high variations were normalized for each combination of stage recovery values. The averages and standard deviations of the 28 normalized values for a specific circuit and a specific indicator were calculated. The standard deviations were usually small, indicating that the indicators do not undergo large variations despite changes in the stage recovery values. Usually, the circuit with the best results has a low standard deviation, i.e., these circuits give the best results independent of the value of the stage recovery. Circuits with moderate results sometimes have significant variation, i.e., the position within the set of alternatives has greater variability. Despite the variation, the values never exceed the values of the best circuits; therefore, these circuits will never be selected based on this indicator.
All of these previous studies provide evidence that the selection of the best flotation circuit is not very sensitive to the flotation stage recoveries, and therefore, only approximate values can be used to select the best flotation circuit. This paper systematically analyzes the effect of the values of the recoveries of flotation stages on the selection of the final circuit.
The oxygen intake of various substances dependent on the storage time is presented as a condition for the CO-formation in Fig. 1 for periods of up to 30 years. Carbon monoxide developed continuously at all relative humidities between 0 and 97 % from graphites, 48 coals, brown coals, peat, wood and some other organic substances (Fig. 2). At some tests under normal conditions the CO-content reached values of more than 0.2 V-% fatal to men, though with some substances no CO-formation was observed at a relative humidity of 100 %, or CO developed which disappeared, however, after some weeks or months. The CO disappeared only when the coal was wet. Reducing the air humidity to 97 % already stopped the CO-transformation. When in a desiccator, in which the CO-transformation had occurred, the CO-free air was exchanged for CO-containing air, then the CO-content of the latter decreased the faster the more often the exchange took place (Fig.3). At last a growing grey spot was observed on the wet coal. When this grey mass had reached a diameter of about 1 cm, it was taken from the coal and suspended in a nutrient solution. With this biosuspension CO could be washed out of the air. Thus it was proved that under normal conditions, CO formed through the oxidation of substances can immediately be transformed by microbes in wet media. Apart from radical reactions, this phenomenon is one reason for the fact that the CO-content of the atmosphere does not increase according to the global emissions of more than 1,000 million t/a. The CO-formation increases with the temperature. It can be proved at 0 C. With coals activation energies of between 45 and 60 kJ/mol CO resulted from the dependence of the temperature for various air humidities. The CO-formation also increases with the oxygen content of the atmosphere.
With almost all substances the lowest CO-formation was observed at medium air and coal humidities. In most cases, especially with coals, the CO-formation was stronger at 0 % r.h. than at 97 %. CO also developed from commercial graphite and from graphite from the flotation plant. Since it seemed possible that this CO came from impurities in the graphite, a graphite sample was heated under exclusion of air to 1100 C for 7 days and then stored in desiccators in air. During a storing time of 8 years the CO-content in dry air remained under 0.001 V-%, which means that in spite of an extremely long storage period a CO-formation could not be proved with certainty. Compared to this, some coals in the same test performance developed a CO-content of more than 0.002 V-% after one day already.
Graphite fresh from the mine, graphite roasted at 1100 C for 7 days, and 7 coals of varying contents of volatile matter were stored in air at 97 % r.h. for 486 days. Fig. 4 shows that the the CO-formation increases with the content of volatile matter. This indicates an increasing stabilization of the coal molecules in the course of carbonization.
Soluble iron salts expedite the CO-formation from organic substances (2). Since pyrite oxidizes to soluble iron salts in wet coal, especially in the presence of the Thiobacillus ferrooxidans (3), it can be expected that it also expedites the formation of of carbon monoxide from organic substances. Indeed, in a period of 24 years in humid air 5 g of humic acid yielded 0.087 and a mixture of 5 g of humic acid and 0.5 g of pyrite 0.139 g CO/100 g of humic acid, which is 74 % more. The assistance of the oxidation through soluble iron salts in the presence of Thiobacilli has been expressed especially clearly in the paper by Guntermann et al (4). In this case the content of volatile matter in the coals changes much more during the desulphurization by iron salts and Thiobacilli than during the storage of wet coal in air.
Basically it is possible that carbon monoxide develops through oxidation from the carbon atoms of the ring structures and/or from non-aromatic groups of the coals. As shown before, only with one graphite heated before storing a CO-formation does not occur. This means that no carbon monoxide develops from highly condensed ring systems. When storing simple purely organic compounds like formaldehyde, formic acid, oxalic acid and acetone, only very small amounts of CO developed in air, from formic acid e.g. only 0.001 % in 61 days. When the substances, however, were stored in the presence of coal acting as a catalyst, they were transformed quickly, especially to CO2 and water. Only the formic acid formed remarkable amounts of CO. In dry air up to 21 V-% CO were measured (1). It was therefore tried to take formic acid as a reaction product of the oxidation of coal. For this purpose fresh and stored coals were degased at 0.5 Pa and 90 C. The pH-value of the condensate from the volatile products decreased in dependence of the storage time. In the condensate from a fresh coal with 26.0 % v.m. (i.waf) it was 5.2, after 7 days storage 3.5 and after 73 days storage 3.1. The acid character of the condensate was caused above all by acetic acid (26 g/100 g coal) and surprisingly by lactic acid (10 13 g/100 g coal), but also by formic acid (0.6 1.0 g/100 g coal) (5). The yield of formic acid is very small; yet it has been shown before that high contents could not be expected since formic acid disintegrates in the presence of coal because of its catalytic effect.
As any other observer, a data reconciliation procedure makes a simultaneous use of measured values and process models. Models are required to cope with common data processing problems in mineral processing: measurement uncertainties, which are quite large in a metallurgical operation, lack of measurement availability for critical variables, limited knowledge of process behaviour, and information redundancy including available measurements and prior process knowledge. The uncertainties about process models are so large in metallurgical industries that it is common practice to use non-causal sub-models such as mass conservation constraints, because the level of confidence in these sub-models is high, and thus prevents distorting the information content of the data sets by complete but uncertain models. In the metallurgical and chemical industries, these observations methods are named reconciliation methods in the sense that they reconcile the measurement data with the laws of mass conservation .
Fig. 7 summarizes the concept. One can see that the procedure core is a least-squares based algorithm that serves simultaneously as an estimator and a filter. It is an observer in the general sense that it allows the observation of process states, an estimator in the sense that it estimates numerical values of state variables which may not be measured or measurable, and finally a filter in the sense that, when a state variable is measured, the observer is able to correct the experimental value of the process state variables.
The input information as seen above consists of experimental data, of their accuracy (measurement error variance), and of the mass conservation constraints. The mass conservation equations are written for all the species involved in a MP plant network: ore, water, metals, particle size classes, metals in particle size classes, reagents, etc. For each plant node and each component or property i that is conserved, the constraint is written (FR=flow rate):
These constraints are linear or multi-linear equations where the parameters are perfectly known since they rely only on the plant flow network and do not take into account transfer parameters such as reaction rates. In steady state data reconciliation, the rates of accumulation are constrained to converge to zero, while for small stochastic deviations to steady-state (stationary operating conditions) the rates of accumulation variances can be appropriately set to represent the variations around the steady-state regime. During transient regimes, the rate of accumulation must be explicitly considered, thus requiring additional modelling information and/or inventory measurements.
The reconciliation algorithm outputs are the corrected measured process variables as well as the estimated unmeasured states (see Fig. 7). Furthermore, it is possible to analytically calculate, or simulate by Monte-Carlo techniques, the variances of the reconciled process variables, therefore allowing to assess the reconciliation results accuracy. Reconciled states are necessarily more reliable than raw measurements, and improve, in this manner, the performances of the subsequent controller and optimizer actions. Let us now look at three examples that will illustrate the benefits brought by data reconciliation for calculating plant performance in steady-state or dynamic state, and for designing instrumentation strategies:
Flotation plant recovery: a copper flotation plant, fed with CuPb concentrates, produces a copper concentrate and a tail which is subsequently fed to a bulk flotation plant delivering a lead concentrate and a zinc-lead concentrate . The traditional 2-product formula can be used to calculate the recoveries from the raw data. Instead, it is strongly recommended to apply data reconciliation before recovery estimation. Table 1 compares the values given by the two methods. The improvement of the recovery standard deviations using reconciled data is now quite significant.
Oil extraction plant: a conditioning plant and an extraction plant  for oil sands are connected through a 30km pipeline exhibiting roughly a 16-h residence time (see Fig. 8). The objective is to produce a coherent mass balance for 24h period operation. When using steady-state bitumen reconciliation of the whole system, the bitumen inventory in the pipeline behaves as a random walk and is therefore not bound to physically acceptable limits. However, when considering the rate of accumulation dynamics, it is possible, as demonstrated by simulation, to keep track of the actual bitumen inventory.
Instrumentation of a flotation plant: the flotation plant of Fig. 9 is to be instrumented with flowmeters combined with on-stream X-ray fluorescence analyzers . The design objectives are: (1) to obtain a complete stream flowrate and composition obervability; (2) to obtain a degree of redundancy 1, i.e. a possible failure of 1 sensor without losing the complete observability; (3) to use sensors of predefined accuracy; (4) to use three sets of magnetic flowmeters coupled to gamma-ray density gauge; (5) to minimize the uncertainty of the plant performance estimator, i.e. the variance of the operating point in the grade-recovery plan. There are 212 possible configurations that satisfy the four first objectives. The optimal one that complies with objective 5 corresponds to the sensor placement shown in Fig. 9.
The installation works of the floating solar panels are over and the plant will be ready for inauguration soon, said Manoharan P, an assistant executive engineer at the KSEB research and dam safety sub-division, Thariyode, reports TheTimes of India.
The project had commenced last year in March, with the plant being set up in collaboration with Thiruvananthapuram-based Adtech Systems Ltd. According to officials of the private firm, the plant will generate 7.5 lakh units of power annually, which will be fed to the KSEB grid using underwater cables.
We have used high-efficiency solar panels for the project as per KSEB stipulations. Also, we have set up a floating substation on the reservoir to convert the output into 11kV considering economic and safety aspects, said Raveendran T Nair, vice-president (projects) of Adtech Systems Ltd.
He also explained that the floating solar plants, when compared to ground-mounted counterparts, showcased higher efficiency owing to the moderating effect of water bodies on panel temperature, besides accumulating lower concentration of dust on the panel.
We at The Better India want to showcase everything that is working in this country. By using the power of constructive journalism, we want to change India one story at a time. If you read us, like us and want this positive news movement to grow, then do consider supporting us via the following buttons:
The Better Home is a range of safe and eco-friendly home cleaners by The Better India. We put out curated content on our Instagram account- green living hacks, green innovations, leading a no-tox life and heroes in the sustainability space.Get in Touch with Mechanic