Search Documents

By D. F. Malan

Considerable practical knowledge and effective use of empirical evidence is required in the design of tunnel support systems. In many instances, it is unclear how individual support elements interact with fractured rock and which elements of the support system are critical in controlling and containing the movement of material proximate to the excavation. The present paper outlines a displacement discontinuity approach that allows the coupling of surface constraints, in the form of liner or mesh material, to an explicit representation of time-dependent rock discontinuity creep movements. The method is illustrated for a simple square-shaped tunnel profile in which the fracture zone is mobilized by an imposed stress field. A crucial aspect of the model is the computational treatment of the coupling between the liner material and there presentation of the fractured rock mass. This allows the response of the liner material to be assessed quantitatively in terms of the possible reduction of movement in the rock mass. The importance of installing the liner timeously after the excavation is formed is illustrated in a specific example. A number of computational difficulties in simulating liner-rock interaction mechanisms are identified.

Jan 1, 2008

By K. I. Afewu, L. C. Woollacott

Little is reported in the literature that is directly useful for the interpolation film coefficients for the in-pulp adsorption circuits used in gold metallurgy. This can create a problem if the design or operation of these circuits is to be optimized by the use of simulation techniques. To address this problem, a study was undertaken on the influence of mixing intensity, scale, and geometry on the film coefficient in adsorption systems typical of the carbon-in pulp plants used for gold extraction. This, the first paper deals with the influence of these factors on the film coefficient and shows that the specific power to just disperse the carbon in the slurry is an important variable. This variable is also strongly dependent on mixing intensity, scale, and geometry-a dependence that is dealt with in the second paper in the series. From the informatlon obtained, practical scale-up procedures were developed. These procedures are considered to be applicable to resin-in-pulp as well as to carbon-In-pulp plants.

Jan 1, 1995

To determine the characteristics of a mineral deposit, the usual practice is to take samples, analyse the properties of those samples and infer the characteristics of the deposit from these properties. This analysis can be done using statistical methods. In the present work we are concerned with two types of statistical approaches, classical statistics and spatial statistics. If one uses classical statistics to represent the properties of the sample values, the assumption is made that the values are realizations of a random variable. The relative positions of the samples are ignored, and it is assumed that all sample values in the mineral deposit have an equal probability of being selected. The likely presence of trends, zones of enrichment, or pay shoots in the mineralization, is ignored. The fact that two samples taken close to each other are more likely to have similar values than if taken far apart is also not taken into consideration. In contrast, spatial statistics will be used if one chooses to consider that the sample values are realizations of random functions. On this hypothesis, the value of a sample is a function of its position in the mineralization of the deposit, and the relative position of the samples is taken into consideration. The similarity between sample values is quantified as a function of the distance between samples and this relationship represents the foundation of spatial statistics. There are few situations where classical statistics can be used. The assumption that all sample values in the mineral deposit have an equal likelihood of being represented will be satisfied only if the sample values are randomly distributed, or if the sample positions are random. Sample values are in fact never randomly distributed within a mineral deposit. Further- more, geologists usually avoid taking samples at random (random sampling), as it is correctly accepted that samples located on a regular grid, or approximately on a regular grid, give more information than randomly located samples. In practice, classical statistics should be used only in the early stages of exploration, when the number of samples available is relatively small and the distances between samples are large. In these circumstances, and whenever the information available is not sufficient to permit the use of spatial statistics, application of the methods described in this chapter is justified.

Jan 1, 1978

By I. Beukes, H. Kotze, B. Krause, R. P. Hattingh

Hillendale Mine is operated by Exxaro KZN Sands, and is situated on the north coast of KwaZulu-Natal in South Africa near the towns of Empangeni and Richards Bay, some 200 km north of Durban. During the mining operation the heavy minerals rutile, ilmenite, zircon and leucoxene are extracted for further beneficiation at a processing complex near the town of Empangeni. The deposit contains approximately 25% slimes (defined as the sub-45 micron fraction), and is mined using hydraulic monitoring. Run of mine is gravity fed down launders to pump stations, which deliver the material to a primary wet plant. Slimes is separated from the run of mine prior to the heavy minerals being separated from the sand. Sand tails containing approximately 2-5% slimes is pumped to the mining void. The slimes is thickened to a density of approximately 1.2, and most of the resultant material is pumped to a sub-aerial deposition site (residue dam). The area is characterized by an annual rainfall of approximately 1300 mm, which is relatively high by South African standards. From an agricultural perspective the soils are deep and well weathered. Dryland agriculture has been practised on an extensive scale in the area for the past 70 years. Sugar cane in particular is cultivated on and around the mine area. In terms of environmental approvals, the post-mining land use and objectives are defined as economically viable and sustainable sugar cane cultivation in the mined areas. During mining the soil profile is destroyed. In this paper the establishment of a reconstituted soil profile is discussed, with specific reference to the methodology that is being followed to achieve the homogeneous blending of sand and slimes. Defining the rehabilitation strategy and soil specification Strategically Exxaro KZN Sands divided rehabilitation into four aspects, comprising physical rehabilitation (restoration of soil structure), vegetative rehabilitation (restoration of vegetative cover), ecological rehabilitation (restoration of sustainable ecological communities), and sustainability (sustainable man/environment interface) (Hattingh and Viljoen, 2006). It is believed that the key to successful rehabilitation lies in achieving a substrate (soil) suitable for the establishment of the target vegetation. During the development of the rehabilitation strategy the key factor in the establishment of a suitable soil profile was shown to be the soil water retention of this material. Since dryland agriculture is practised, this aspect becomes critically important in ensuring that the post-mining land use is sustainable (Hattingh et al., in prep.). From a financial perspective, a closure certificate being issued implies that the state takes over financial liability for the rehabilitated land. This has significant implications for the state-the default position will be that all possible latent and residual impacts must be identified and known prior to a closure certificate being issued. From the mine's perspective this precautionary approach by the state implies that post-operational costs cannot be quantified exactly, and our approach is therefore to mitigate the risk of a long post-closure care and maintenance phase by means of addressing all potential latent defects during the operational rehabilitation phase. The financial viability of the soil reconstitution process itself is therefore not measured against other rehabilitation methods, but against the risk of not achieving closure. During a series of technical workshops comprising experts in the various disciplines in the rehabilitation of the soil profile were identified. These aspects include the re-establishment of a functional soil profile, soil physical characteristics, soil chemistry, soil biota, salinization potential, erodibility and soil weathering. The natural soils in the Hillendale area have been classified in terms of the South African soil classification system (MacVicar and co-workers, 1991) and consist of Hutton (orthic topsoils on red apedal subsoils) and Clovelly (orthic top soils on yellow apedal subsoils) soil forms. The topsoils consistently comprise medium to fine grained sand overlying heavier textured sandy clay loam to clayey subsoils. The topsoil structure consists of a single grain structure, whereas subsoils are apedal. These soils are typically deeply weathered and do not show secondary structures. During the mining process a layer of topsoil (300 mm) is stripped off, and the balance of the material down to the footwall is processed. Sand with 2-5% slimes (defined as the sub-45 micron fraction) is returned as backfill in order to re-establish the dune to approximate the previous landform. From an agricultural perspective this material is not suitable for the dryland cultivation of sugar cane. A capping of more suitable material therefore has to be provided. The most significant difference between the premining soil form and the backfill material is the percentage slimes, consequently a mixture of the sand and slimes is required as a capping over the backfilled sand. The required thickness of the cap is a function of specifically plant growth requirements, erodibility, weathering and runoff (Hattingh, et al.; in prep.). The percentage slimes required in the cap is a function of mainly soil water requirements (Hattingh and Viljoen, 2006). The conservative indications to date show that the reconstituted soil profile should consist of a homogeneously blended mixture of 70%-80% sand and 30%-20% slimes up to a maximum of 2 metres thick. This will be sufficient to ensure growth during dry periods.

Jan 1, 2007

By M. D. G. Salamon

In view of an urgent need for a simple means of simulating the behaviour of stratified rock masses, a long-dormant linear model is revived, generalized, and discussed in some depth. The model is piecewise homogeneous, consisting of a pile of homogeneous isotropic laminae in which the interfaces between beds are parallel and free of shear stresses and cohesion. The continuity of stresses and normal displacements is assured across the bedding planes and, to maintain simplicity, the theory of thin plates is used as the kernel of the model. The paper covers the case where the rock mass consists of discrete beds in which the layers are arranged either heterogeneously or systematically. The stratification is regarded as uniform when all the layers are of the same thickness and their Young's modulus changes exponentially with depth; and as homogeneous when all the laminae have identical thickness and properties. The case of uniform stratification is also treated on the basis of an equivalent quasi-continuum derived from the basic model. Attention is focused on the derivation of influence functions. It is demonstrated that the quasi-continuous models, when they are available, can replace their distinct-layer counterparts. The homogeneous quasi-continuum, which permits the derivation of its influence functions in closed form, facilitates the determination of the roof and floor convergence. In the case of a significant class of problems, the convergence distribution is defined by a second-order differential equation in the xy plane, which simplifies to a Poisson's equation if the mining excavations are unsupported. With this result, the paper prepares for the solution of a host of practical problems. These solutions are presented in a series of papers that are in the course of publication.

Jan 1, 1991

By F. Lombard

We consider in this paper the use of a modified version of Hotelling?s statistic in the analysis of particle size distributions. The statistic can be adversely affected by the presence of outliers among the data. We propose a competitor to the statistic that is based on ranks, and hence is less sensitive to outlier effects. The results of a Monte Carlo study suggest that the rank test is highly competitive with the Hotelling test in its ability to detect differences between two mass size distributions. The calculation of the rank statistic is explained in detail and its application is illustrated on two sets of data.

Jan 1, 2012

By F. Lombard

We consider in this paper the use of a modified version of Hotelling?s statistic in the analysis of particle size distributions. The statistic can be adversely affected by the presence of outliers among the data. We propose a competitor to the statistic that is based on ranks, and hence is less sensitive to outlier effects. The results of a Monte Carlo study suggest that the rank test is highly competitive with the Hotelling test in its ability to detect differences between two mass size distributions. The calculation of the rank statistic is explained in detail and its application is illustrated on two sets of data.

Jan 1, 2012

This paper summarises some observations regarding the way SAG and AG mills work based on the last ten years of pilot plant operation, grinding circuit design and operational reviews for various projects in Australia and overseas. The main purpose is to promote discussion on key factors influencing breakage in a SAG mill. The focus of the paper is on two questions: ? How do mill lifters and liners really work and have an effect on grinding? ? How does the mill charge composition impact on grinding performance? The focus of the paper is on observations of mill behaviour in both pilot plant trials and plant operation, particularly on Cadia ore. DEM, by others, and JKSimMet modelling of SAG mill performance have greatly assisted in forming the ideas expressed in the paper and both are valuable tools in developing our understanding of the comminution process. Put in the `extreme' form the major observations and conclusions expressed in this paper are that: ? the main purpose of mill lifters is to reduce mill liner wear and not to promote rock breakage; ? the often promoted theory that balls impacting the ?toe of the charge? is a dominant form of breakage is incorrect; and ? fines production in a SAG mill is primarily a function of the rock load mass, with a secondary dependence on ball load for any given circuit configuration.

Jan 1, 2007

The mathematical models of semivariograms can be classified according to whether or not they correspond to situations where the second order stationarity conditions are satisfied (§ 3.3.4). If second order stationarity exists, the a priori variance s2 of x (z) also exists and is finite. For large distances h such that x (z) and x (z + h) are not correlated, the semi- variogram ? (h) will reach a value equal to s2 [this is a consequence of (3.27) above]. This limiting value is called the sill C of the semivariogram, and the distance at which ? (h) reaches the sill is called the range. The range of the semi- variogram obtained from a given type of sample corresponds to the distance of influence of these samples. If only the intrinsic hypothesis is satisfied, the sample variance is an increasing function of the size of the area or volume in which the samples are located, and the semivariogram ? (h) will be a continuously increasing function of h. Remark: It will be shown in Chapter 7 that the variance s2 of x (z) in the ore body O is equal to the average value of ? (h) in O. If ? (h) is a non-decreasing function with sill C, the average value of ? (h) in O will be less than C. s2 and C will be equal only if for some values of h in O, ? (h) exceeds C. In practice, if the range of the semivariogram is small with respect to the dimensions of the ore body, the difference between C and s2 will be negligible.

Jan 1, 1978

By D. Moradkhani, S. M. Javad Koleini, M. Abdollahy, F. Soltani

"The objective of the study was to determine the type and distribution of rare earth elements (REEs) in the phosphate flotation concentrate from the Esfordi flotation plant in central Iran, in order to understand their leaching behaviour. Samples of the concentrate were prepared and analysed using XRF, XRD, ICP, and EMPA. The Ce, Nd, La, and Y contents were 5608, 2227, 1959, and 679.7 mg/kg, respectively. The main REEbearing minerals in the concentrate are fluorapatite, monazite, and xenotime. Apatite is the chief mineral of the concentrate and the ratio of apatite to monazite is greater than 60. From EMPA analysis, monazite contains a much higher proportion of REEs than apatite, but more than 50% of the REE content of the concentrate is in apatite, as this is the main mineral constituent of the concentrate. During the leaching of REEs in the phosphoric acid production process, more than 99% of the total rare earth values report to the large amounts of phosphogypsum (CaSO4·2H2O) formed during the dissolution reaction. Accordingly, two process options are suggested for treating the Esfordi concentrate. In the first option, the concentrate is treated with 98% H2SO4 at 200–300°C. After digestion and water leaching, phosphorus and REEs are precipitated out of solution. In the second option, the concentrate is treated in a pre-leach stage. The process residue from the pre-leach stage is then subjected to acid digestion at elevated temperature for phosphoric acid production.IntroductionThe rare earth elements (REEs) comprise a set of 17 chemical elements in the periodic table, specifically the 15 lanthanides plus scandium and yttrium. REEs and alloys that contain them are used in devices such as computer memory, rechargeable batteries, cell phones, catalytic converters, magnets, fluorescent lighting, and many more (Krishnamurthy and Gupta, 2004). China supplies about 94% of the REE demand, with the remaining 6% coming from Russia and Estonia, the USA, India, Malaysia, and Brazil (Zhanheng, 2011). Increased industrial development in China has prompted the Chinese government to limit annual export quotas to approximately 35 kt of rare earth oxides (REOs), while non-Chinese annual demand is expected to reach 80 kt by the year 2015. This constriction of supply is being met by the development of many new rare earth mining projects, each of which has its own unique mining and processing challenges (Jordens, Cheng, and Waters, 2013). One of these projects is the Esfordi phosphate project, which is located in Yazd Province in central Iran. The concentrate from the Esfordi flotation plant contains 1.2–1.5% REEs."

Jan 1, 2017

By P. C. Schutte

There have been several high-profile accidents in the South African mining industry where fatigue was identified as either causal or contributory. The risk of fatigue is inherent in any work-time arrangement involving shift work and work that is physically or mentally demanding, repetitive, or requires high vigilance. The management of fatigue is not simply a matter of correct shift scheduling. Amore comprehensive approach is needed. In order to be successful, a fatigue management programme should address the unique needs of the operation in which it will be implemented and should be integrated into normal operations. Fatigue management is a shared responsibility between the organization and its employees, and all stakeholders should participate in order to provide the safest and healthiest working environment possible.

Jan 1, 2010

Statistics is essentially a study of variability, the use of some suitable mathematical model representative of such variability, and the application of this inferred pattern of behaviour to practical problems. Probability theory is the mathematical structure devised for providing models for chance happenings. Perhaps the simplest example of a chance happening is the result obtained in tossing a coin. Although the result of a single toss can never be predicted with certainty, there is no reason why in repeated tosses the expectancy of heads and tails should not be equal. In such a case the outcomes of the experiment are said to be equally likely, so that the probability of obtaining heads (or tails) is 1 in 2, that is, 0,5. This simple probability .model refers to a 'mathematical' or 'unbiased' coin. Whether such probabilities are meaningful for actual coins tossed in any particular way is a matter for experience to determine. Probability models have also provided the basis for representing observed variability in cases where the actual under- lying probabilities are only partly known or defy analysis. A typical example of the latter case is the variability observed in the values obtained from gold or other ore samples. The interacting natural forces which gave rise to this variability are obviously very complex and are unlikely ever to be unravelled completely. In spite of this, probability theory and the more recent geostatistical theories have provided us with models which can represent this variability on a reasonable basis. Applied statistics and geostatistics must therefore never be accepted as exact mathematical sciences. Furthermore, a basic and essential concept is that in the type of applications covered in these notes, no model must ever be accepted as necessarily the final and exact answer to the problem; thus, any model must prove and continue to prove its appropriateness in practice, preferably by follow-up comparisons.

Jan 1, 1981

By J. A. Voster

Premier Mine (a part of De Beers Consolidated Mines Ltd) which is situated 38 km to the east of Pretoria, has the largest known diamond pipe in South Africa. The mine started operations in1903. The Premier Mine kimberlite pipe is unique in that it is intersected by a 75 m thick gabbro sill at a depth of 400 m below surface. The current mining method used is mechanized panel retreat caving, which has been successfully used to extract ore from the different mining blocks. All caving at Premier Mine is below the sill and failure to exercise good draw control could cause the early ingress of gabbro waste. Draw control at Premier Mine is crucial as it aids in preventing conditions of isolated draw, which results in the early ingress of waste, compaction of the kimberlite and the loss of ore reserves. To assist, an underground vehicle monitoring and control (VMAC) system was installed to enable the effective management of the draw control function. The system has evolved graduallyin collaboration with a local developer. The data accumulated is used to decide the most effectivedraw point loading pattern in order to achieve sound draw control principles. This is achieved as each load can be related back to the draw point of origin. The VMAC data recorded also relates back to specific load-haul-dumps (LHDs), tunnels, ore passes, as well as indications of dates, times and payloads recorded. An interface exists between the VMAC and a draw control linear program. Premier Mine’s VMAC system is not a ‘dispatch’ system but rather a draw control system that caters for Premier Mine’s specific block cave requirements. Premier Mine has realized that successful automation is not dependent on technology alone but rather on the successful implementation thereof.

Sep 1, 2001

By J. Ndlovu, R. Hundermark

Keywords: Anglo Platinum, Polokwane Smelter, pyrometallurgy, metallurgy, platinum, platinum group metals, PGM, base metals, nickel, copper, sulphide smelting, electric furnace, Söderberg, electrode, matte, slag, flash drying, matte casting, crushing, granulation Anglo Platinum?s Polokwane Smelter is situated outside Polokwane, in the Limpopo Province of South Africa. Wet concentrate is received from various concentrators along the Eastern Bushveld Complex, with 60% of the total concentrate received being from UG2 reef and 40% from Merensky reef. The concentrate grade differs from 35 g/t to 110 g/t, and from this a PGM-rich nickel-copper matte is produced. The furnace matte produced is sent to the ACP converter in Rustenburg. The Anglo Platinum concentrators, from which concentrate is received, are Lebowa, Potgietersrus, Amandelbult, and Union, as well as the joint venture Modikwa. The concentrate is fed to two flash dryers that utilise coal-fired, fluidized-bed hot-gas generators to produce the hot gas that will drive off the moisture, leaving a fine bone-dry concentrate to feed to the furnace. The dry concentrate is pneumatically transferred to a storage silo before being transferred to the feed bins situated above the furnace. Lime, if required as a flux, may be transferred separately to the furnace feed bin by a pneumatic system. The concentrate and lime is fed from the feed bins onto airslides and into the furnace. The feed rate is automatically controlled, and the system is set up to optimize the power to feed ratio. The electric furnace is nominally rated at 68 MW and the power is transferred into the furnace using six 1.6 m diameter Söderberg electrodes. Concentrate is melted by energy generated when electric current passes through the electrodes and resistive slag layer. On melting, two immiscible phases form: slag and matte. Furnace matte, containing the bulk of the base metal sulphides and PGMs, is denser than slag and collects naturally at the bottom of the furnace. The furnace is constructed of a combination of refractory brick and water-cooled copper coolers. The furnace sidewalls and hearth are cooled by air drawn through the area by cooling fans. The hearth and matte zone are constructed from refractory bricks. The copper coolers reside only in the slag zone of the furnace along the entire perimeter of the furnace. One staggered row of plate coolers is installed above the waffle coolers along the perimeter of the furnace. Matte is tapped periodically through one of the two matte tapholes into 35-ton refractory-lined ladles, and cast into matte ingots on a casting machine. The cooled matte is discharged onto a concrete slab for cooling and then transferred by front-end loaders to the crushing plant. The matte is first crushed in a jaw crusher and then a Rhodax cone crusher to a size of 2 mm before it is loaded into tankers and sent to the ACP. The low-grade slag is tapped from the furnace through a water-cooled copper insert into a short water-cooled copper ?hot launder? which discharges into a granulation cold launder. The granulated slag slurry flows to three rake classifiers, from which the coarse slag is discharged onto a conveyor to dewatering silos. The dewatered slag from the silos is dumped. The water overflow from the classifiers reports to two thickeners where the water is clarified, and the thickener underflow is returned to the rake classifiers. The smelter is a zero-effluent plant and all the water from the slag silos where the slag is dewatered, as well as the storm water, is pumped to the process water dam from where it is pumped back into the plant for process water use. The off-gas from the furnace is drawn through a forced draught cooler and into a high temperature baghouse. The dust collected in the baghouse is pneumatically transferred into bins above the furnace and is fed via the concentrate airslides. The cleaned off-gas is then vented through the main stack.

Jan 1, 2006

By J. A. C. Diering

All underground openings are three-dimensional, and in many cases it is not valid to analyse two-dimensional sections. This has been a problem in the past owing to the magnitude and complexity of three-dimensional stress analysis. In this paper a three-dimensional boundary element approach, using non-conforming quadratic elements, which is specifically applicable to underground excavations, is described. In addition, the philosophy of application of stress analyses is discussed. Three case histories are presented to demonstrate the practical nature of the approach as a tool in solving mining problems.

Jan 1, 1987

In many practical situations, calculation of the point semivariogram is not necessary. Such a case is analysed in this chapter. The purpose of this analysis is to give a better understanding of the meaning of the mathematical formulae developed in the previous chapters. A similar problem will be studied in Chapter 11, using the point semivariogram.

Jan 1, 1978

By K Kongolo, M K. Kalenga

It is well known that the electrochemical precipitation of metals from aqueous solutions is conducted by the electrochemical standard potential scale. This scale indicates the order of nobility of different metals. Consequently, the standard potential scale shows the nocivity of metals when they are present as impurities in aqueous solutions containing metals. In the hydrometallurgy of cobalt, nickel and zinc are co-deposited during cobalt electrodeposition. Nickel has almost the same chemical, physical and electrochemical properties as cobalt. Furthermore, the electrochemical potential of nickel is relatively higher than that of cobalt. For this reason nickel is preferentially reduced at the cathode in the presence of cobalt. The presence of small amounts of zinc in solution leads to contamination of cobalt electro won. In spite of its lower electrochemical potential in comparison to cobalt, zinc is reduced on the cathode because of its higher hydrogen-overvoltage. But, in the hydrometallurgy of zinc, small amounts of cobalt present into solution will contaminates the zinc reduced at the cathode. This is easily explained by the fact that cobalt is more electropositive than zinc. The co-deposition mechanism of sulphur at the cathode is not yet well understood. Great attention has to be paid to the mechanism of co-deposition of sulphur during cobalt electrolysis..This work aimed to investigate the influence of some parameters on the co-deposition of nickel, zinc and sulphur during cobalt electrowinning. The following parameters have been investigated : Co2+ concentration in the electrolyte, current density and ammonium sulphate (NH4)2SO4 additions. It has been established that important nickel, zinc and sulphur co-deposition takes place at relatively low Co2+ concentration in solution. The behaviour of these elements with regard to the cathodic phenomena has been established.

Jan 1, 2005

By W. Bleloch

Published in the Journal, September 1970, and presented at a colloquium on 'Future of the Steel Industry', 17th March, 1971. Errata (a) The final sentence in the penultimate paragraph, left hand column p. 34 of the text should read: 'In the hot regions of the bosh and stack the C+CO2=2CO reaction proceeds to CO from coke carbon and carbon dioxide generated by reduction of iron oxides by CO (the so-called solution loss of carbon)'. The italicised words have been omitted in the text. (b) Temperature in Fig. 3 is K not C. (c) In Fig. 9, connection from top plate of column 10 to heat exchanger 9 has been omitted. DISCUSSION J. B. BEEDON AND C. J. JONKER* Dr Bleloch's paper has outlined the advantages of recycling stack-gas of low nitrogen content in the blast furnace. In the first instance it would appear to be extremely beneficial in controlling the hearth temperature and allowing a 100 per cent oxygen-blast to be used. On a second point however we have to disagree. It is claimed that heat is liberated in the combustion zones by the oxidation of CO to CO2 near the tuyeres and in the subsequent reduction of the CO2 by CO. The author has hypothesized as follows, and we quote: 'The C+CO2=2 CO reaction in the outer combustion zone at about 1 700°C is no longer endothermic due to loss of stability of CO2 and its consequent dissociation into CO and free oxygen.' This statement requires closer examination. If we consider the individual events occurring with the CO in the recycled topgas, with the oxygen in the blast and with the coke in the burden, as done by Dr Bleloch, the following heat considerations emerge. All heats of reaction are calculated from the enthalpy values of each reactant and product, as well as the temperature dependence of their molar heat capacities. (a) At the tuyeres: CO+1/2 O2-+CO2 . . . . . (1) CO, CO2 at 500°C; O2 at 25°C. ?H reaction = -65 000 cals/mole CO The reaction (1) is exothermic. (b) In the outer combustion zone: CO2+C?2 CO . . . . . . . . (2) C, CO at 1 700°C; CO2 at 500 0C. ?H reaction = -56 000 cals/mole CO Reaction (2) is endothermic. The overall heat gain of reactions (1) and (2) is ?H=+56 000 - 65 000= -9000 cals/mole CO i.e. the overall reaction is slightly exothermic, the amount of heat liberated is however smaller than suggested by Dr Bleloch. (c) The dissociation CO2+CO+1/2 O2 . . . . . (3) is a highly endothermic reaction at 1 700 0C, requiring 82 000 cals/mole CO2 to proceed to completion. Furthermore, this reaction (3) is energetically not favoured to proceed, as can be observed from the free energy-temperature relationship of figure (3), which favours the formation of CO2 at 1 700 0C. Hence direct dissociation of CO2 at 1 700 0C appears to be unlikely.* The concept of recycling topgas is however still intriguing from many other points of view. Apart from providing a very efficient method of hearth temperature control, it will create a much higher CO partial pressure in the stack. Whether the rate of indirect reduction in the stack will be proportional to this partial pressure or not is un-

Jan 9, 1970

By G. T. van Rooyen

Materials for Winding Plant Components By Prof. G. T. van Rooyen (ScD) (MIT). Materials for critical applications such as winding plant components are reviewed in terms of the basic requirements which have to be complied with. It is shown that tensile testing alone is not sufficient and that impact testing also is necessary if the incidence of brittle fracture is to be avoided. As a criterion it is suggested that the brittle transition temperature (B.T.T.) determined by impact testing should in all cases be well below the lowest anticipated service temperature. A number of steels were evaluated by impact testing and it is shown that a small ferrite grain size, obtained either by chemical means such as by the use of appropriate alloying elements in the steel or by means of a quenching heat treatment, is effective in attaining a Iow B.T.T. The speed of strain aging as well as the influence of strain aging on the B.T.T. was also evaluated. On the basis of the results obtained it is recommended that steels for critical applications must conform to an appropriate grade and class of steel according to BS 4360 specification. It is further recommended that such steels be exempted from the customary six-monthly heat treatment prescribed by the Mines and Work Act, Regulation 16,18.

Solvent loading: It is agreed that it is advantageous, as far as reagent cost is concerned, to operate at maximum solvent loading. However, owing to the fact that no reliable automatic solvent flow controller is yet available, control tends to be erratic with the result that there is risk of losing uranium. As can be seen from Mr Tunley's plot, the graph representing total ammonia cost is fairly flat between loadings of 3.5 and 4.0 gpl, while the graph representing uranium loss in the raffinate is steep. It is, therefore, safer to operate slightly below optimum loading in the absence of absolute control. Should the purity of the final product gain importance, operation at optimum loading or even slightly above, may be advantageous. Solvent losses and clarification of pregnant solution: The cost of replacing solvent, as indicated by Mr Carman, represents a high percentage of the overall Purlex costs. It is, therefore, agreed that, amongst other factors, efficient clarification of the pregnant solution and mixer-settler design have an important bearing on the cost. However, as at Blyvooruitzicht, the Buffelsfontein plant can tolerate a suspended solids content of 50 ppm without serious deterioration in efficiency. An increase in the suspended solids content from 15 to 50 ppm normally accounts for an increase of approximately 1 ppm in amine loss. Clarification of acidic pregnant solutions is much more difficult than that of alkaline solutions. Sand clarifiers are being used at Buffelsfontein with moderate success. Clarified solutions containing less than 20 ppm of solids in suspension can be produced, depending on the quantity and type of flocculant used in the filtration process, the thickness and the grading analysis of the sand beds and the quality of supervision. An on-stream duty of approximately 0.15 gallon per sq ft of surface area is obtained. Hydraulic handling of clarifier sand, while cleaning, may effect the efficiency of a clarifier, as size segregation of sand particles may lead to inefficient "pockets" in a sand bed. Control of the solution level in the sand clarifier also affects the quality of the effluent. Exposure of the sand bed invariably causes a cloudy effluent. Recycling of the final and initial solution discharged before and after skimming the clarifier sand improves the clarification efficiency significantly. Post-precipitation of silica and possibly sulphate salts, after clarification, may give a false impression of inefficient clarification. This phenomenon is more pronounced when treating hot supersaturated solutions. Ageing and cooling of the solution before clarification assist in stabilising the solution and reduce precipitation after clarification. The maintenance cost of sand clarifiers at Buffelsfontein is in the order of 0.35 cent per metric ton of solution treated. A redundant ion-exchange column has been tested successfully, either as a primary or a secondary clarifier, depending on the quality of the filtrate treated. The conventional pebble and sand beds in the column have been retained. Operation was enhanced by the addition of a two-foot deep resin bed as a filtering medium. The suspended solids content was reduced to less than 5 ppm. Mixer-settlers: Mixer-settler design is a controversial subject but, in view of the high costs of solvent compared to relatively low pumping costs, it is believed that mixing and pumping functions should be separated. Separate facilities for pumping and mixing not only give greater operational flexibility but also permit eventual changes in the flow circuit as technology develops. Scrub technique: Unlike the pilot plant, where the crud was carried away in the aqueous phase, crud is carried over with the organic phase in the full-scale plant. This carry-over of crud into the strip bank aggravates crud formation in the first strip stage. This phenomenon may result from a slight difference in settler design and/or a variation in the relative phase depths in the two plants. Bypassing of the scrub aqueous solution from the fourth to the second stage was introduced in order to avoid dilution and partial neutralisation of the acid added to the third stage, as it is known that a low pH value enhances the transfer of iron into the aqueous phase. The low aqueous flowrate in the third scrubbing stage was overcome by introducing an aqueous phase recycle. Gleaning: Mr MacDonald's contribution on "gleaning" was most informative. This process for recovering solvent from raffinate seems to show at least as much promise as coalescers and other methods employed to date. Similar tests conducted at Buffelsfontein pilot plant gave 60 per cent amine recovery from raffinate containing 10 ppm amine. However, the circuit was not operated at equilibrium and recovery may be expected to drop as the amine concentration of the "Gleaner" circuit increases.