Search Documents

By J. D. Steenkamp

Keywords: Thermal, conductivity, diffusivity, expansion, history, density, specific heat capacity, industrial rich high-carbon manganese slag The thermal conductivity of slag is an important parameter in the design of furnace-containment systems based on freeze-lining technology. Literature indicates that the crystal content of a slag has a significant influence on its thermal conductivity. Industrial, rich, manganese-bearing slag was cooled at different rates to create samples with different microstructures. The coefficients of thermal conductivity of these samples were measured in a nitrogen atmosphere from room temperature to 990°C at 100°C intervals. The laser-flash measurement technique was used for this purpose. Follow-up investigations included XRF and XRD and modelling in FACTSage.

Jan 1, 2011

By R. W. Pugh, D. G. Bellow

"This paper will describe how cultivator shanks are made in one company and how the manufacturing was studied with a view to increasing product consistency and reducing manufacturing cost. The product: which is used extensively in the agriculture industry, was analyzed to determine which geometric parameters optimized its design and whether there was justification for the great variety of styles and shapes that are available in the market. Finally, an extensive series of field trials were undertaken which showed that it is advantageous to the manufacturer to be better aware of the loads and forces that his product is subjected to by another industry.IntroductionStelco maintains a Research Group for its Western Region operations headquartered in Edmonton, Alberta. Because of the small staff, a location that is isolated with respect to its Burlington, Ontario centre and a diverse set of processes and products, a university relationship has developed which may be different from that which would be developed at a larger integrated Steel Works. This paper outlines the type of cooperative studies carried out by Stelco Inc. in Edmonton and the University of Alberta. .To meet the research needs covering a wide variety of steelmaking processes and steel products produced in Edmonton would require more staff and equipment than the sales volume in the West justifies. The gap is filled by cooperation with university personnel. This cooperation has provided these benefits to Stelco Inc.:• access to specialized test equipment and a trained staff to operate it; and• assistance from experts in the areas of product testing and problem solving.Moreover, the close proximity of the university to the plant allows frequent meetings and ensures that practical solutions are developed and implemented. A good example of the ongoing university I industry relationship developed is the joint work on improving cultivator shanks."

Jan 1, 1982

By G J. Lyman, F S. Bourgeois

As the prices of metals slide, financial control of mineral processing operations becomes more and more important. Control of a mineral processing operation implies measurement of grades, recoveries and losses, with every effort being made to recover more product from the same feed. Shareholders become more concerned with the transparency of the operations and the corporate governance.There is only one way to achieve this and this is by sampling and making good use of the results of that sampling. Sampling systems that provide only 'rough' figures are not good enough. Similarly, a competitive position in the marketplace demands close control of product quality. Maximisation of product output demands that the product just meet the specification so that product yield is maximised. Decisions based on precise sampling results carry less risk that those based on imprecise results.This paper reports some recent advances in sampling theory that permit full quantification of sampling distributions, not just the sampling variance as has been possible in the past. These advances can be used together with simulation methods to tie down sampling precisions and develop calculations that quantify decision making risk when using the sampling results. A number of examples illustrate the use of the new advances and a final example indicates the development of risk analysis curves in the coupling of sampling precision to sale of a commodity.CITATION:Lyman, G J and Bourgeois, F S, 2015. Sampling, corporate governance and risk analysis, in Proceedings MetPlant 2015, pp 27–34 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Sep 7, 2015

By W Barry Gregg

The rock blasting and mining industries use a system of detonators and explosives for mining operations. Many detonators use small metal bridgewires, which are heated by passing a current through the wire, heating and igniting the pyrotechnic powder in contact with the wire, Typical bridgewires require a current of at least 3 amps, applied for a few milliseconds, for an explosive output. The semiconductor bridge (SCB), in comparison, initiates an explosive in 0.01 to 0.1 of a millisecond.

Jan 1, 1992

By David A. Larson

Preliminary laboratory experiments have been conducted by the Bureau of Mines to define the influence of bit geometry on methane ignitions as a step toward establishing safer mining-machine operating parameters. The experimental system was designed to simulate the operation of a continuous mining machine, using a drum-type cutterhead of normal diameter with a narrow cross section, permitting full-diameter drum cutting tests to be run on material representative of sandstone lenses (or inclusions) in coal seams. Two conical bit types were tested at shallow depths of cut with three drum speeds. Bits that rotated in the mounting block during testing did not produce ignitions. The ignition frequency for used bits was three to four times greater than for new bits at the same kerf depth.

Jan 1, 1983

By M. M. Tilman

Etchants containing hexavalent chromium and sulfuric acid are used in a variety of surface-finishing operations. When the resultant spent solutions are discarded, substantial quantities of chromium are lost and pollution prob¬lems are created. To minimize the undesirable environmental effects while improving metal and minerals recycling technology, the Bureau of Mines is con¬ducting research on a method for the economic inplant recycling of these waste etchants. Trivalent chromium, produced when the etchant reacts with a substrate, is oxidized in the anode chamber of a diaphragm cell; other metals, dissolved during the etching operation, are transferred to the catholyte. When waste brass etchants are treated in this manner, all of the trivalent chromium is oxidized and more than 40 percent of the copper and zinc is removed. Similar copper removal and chromium oxidation are achieved with waste printed circuit board etchants and rinse waters from plastic etching operations. The electrical energy needed to regenerate 1 kg sodium dichromate is less than 9 kwhr. Regenerated brass etchants evaluated by two companies equaled or exceeded the performance of fresh etchants.

Jan 1, 1979

By Lawrie D. C

Finely ground beryl was decomposed by mixing with ground limestone and soda ash and heating at 900°C for 4-6 hr. The sintered mater:al was then crushed and leached by boiling under reflux with 5N sulphuric acid for up to 4 hr. The liquor containing Be, AI, Fe and other elements as sulphates, was separated from the residue which consisted mainly of silica and gypsum, either by filtration or bydecantation washing.The clarified leach liquor was first treated to remove Fe+++ by contacting with a primary amine-kerosene solution in 2 countercurrent stages. The iron free liquor was then passed to a 4-stage counter-current extraction unit where it was contacted with O·5M di- (2 ethyl hexyl) phosphoric acid (EHPA) in kerosene. Under the operating conditions, 99 per cent of the Be and 7 per cent of the Al was extracted.The Be and Al were stripped from the organic phase by sodium hydroxide solution of suitable concentration to form soluble "beryllate" and "aluminate" respectively. Most of the Be was subsequently precipitatedas Be(OH)2 by heating the alkaline solution to 90°C, the Al remaining in solution. After filtration and washing, the Be(OH)2 could be ignited directly to BeO at about 1000°C or used as a feed to further purification stages if required.A small counter-current extraction unit was operated continuously using liquor derived from the sintering and leaching stages. Dataobtained in batch laboratory tests were confirmed. A product obtained by ignition of the washed "beryllate" precipitate contained more than 99 per cent BeO.INTRODUCTIONIn 1958 an investigation was commenced into the recovery of beryllium oxide from beryl, as part of a programme to utilize mineral resources of South Australia. A survey of the literature indicated that little work had been done on direct solvent extraction of beryllium from aqueous liquors, so it was decided to apply the experience gained in this field in uranium processing to the current problem.Preliminary extraction tests on solutions of beryllium...

Jan 1, 1965

By John W. Sweeney

The southeastern phosphate industry is investigated to determine its role in supplying the United States and world phosphate demand. The report shows that the southeastern (Florida, Tennessee, and North Carolina) phosphate industry is the leading domestic and world producer of phosphate rock, The report emphasizes size and organization, production capacity, transportation, and marketing of the southeastern phosphate product, Factors affecting the supply- demand relationship of the southeastern phosphate industry and world phosphate production-consumption relationships are analyzed and related to the southeastern phosphate industry. Past trends are projected and the future role of the southeastern phosphate industry is analyzed to determine its ability to supply the United States and rest-of-the-world phosphate requirements in 1985 and 2000. The projections show that southeastern resources are ample to supply domestic and export requirements at the projected rates of consumption well beyond the year 2000.

Jan 1, 1970

By J. Jeffrey Wagner, Michael J. Bailey, Peter M. Douglass

The Chambers Creek Interceptor Sewer Tunnel project consisted of 2-1/2 miles (4 kilometres) of nominal 9-foot (2.7-metre) O.D. tunnel driven through predominantly glacially overridden soils. Mole-Collucio Construction Company initially elected to drive the entire tunnel with a Lovat TBM. However, this machine could not control the fast raveling ground conditions which developed when the TBM encountered sands overlying very dense, hard glacial till or boulders. As a result, a shield and digger machine was brought in for a 1700-foot (518-metre) long segment. Pumped wells spaced roughly 50 to 125 feet (15 to 38 metres) apart were utilized to dewater about two-thirds of the alignment during construction. Initial support consisted of steel ribs and wood lagging, with a final lining comprised of pre-cast concrete pipe segments grouted in place.

Jan 1, 1985

As far as is known at the present time, important ore occurrences at Mount Isa are restricted to the Urquhart Shale. These occur at Mount Isa itself, and some 13 miles north along strike at the Northern Leases.The primary silver-lead-zinc and copper orebodies occur as adjacent but discretc entities in lithologieally distinct host rocks. The copper orebodies occur in an envelope of "Silica Dolomite", whereas the silver-lead zinc orebodies occur in shale typical of the Urquhart Shale in general.At both Mount Isa and the Northern Leases, secondary lead orebodies occur at thc surface representing primary stratiform silver-lead-zinc sulphide accumulations in the fresh rocks at depth.Secondary copper orebodies are known only at Mount Isa, where they have formed in favourable environments, but unlike the secondary lead orebodies, not in the positions of the former primary mineralization that gave rise to them. The secondary copper is localized in shale remote from the point where the copper was released into solution by weathering. It was derived from very low grade fracture fillings and disseminations of chalcopyrite in the outcropping or suboutcropping crest of the "Silica Dolomite".Primary silver-lead-zinc orebodies arc oxidized without significant transport or secondary enrichment, but secondary copper orebodies arc characterized by considerable transport and extreme secondary enrichment.The original water table stood at approximately 1,150 ft (mine datum) and for the majority of the mine area controlled the base of complete oxidation fairly closely.An oxidized zone approximately 175 ft deep is general in the Urquhart Shale.Structures affecting ground conditions have influenced oxidation and leaching of the silver-lcad-zinc orebodies, and have had effects vital to the formation of secondary copper orebodies.

Jan 1, 1966

By T Christie, A Swain

Patterns of quartz veining in epithermal gold-silver deposits at Wharekirauponga, Neavesville and Ohui in the Hauraki Goldfield, have been analysed to determine the effects of different lithologies on vein distributions. At Wharekirauponga, quartz veins are hosted in rhyolite and minor andesite, with the andesite containing slightly higher gold grades. Local basement andesite, rhyolitic tuff and silicified black shale host gold-bearing quartz veins at Neavesville. At Ohui, quartz veins are hosted in rhyolite, rhyolitic tuff and andesite. Fractal analysis was used to quantitatively describe the distribution of veining. The fractal dimension (D) of veining in silicified tuff is 0.25, moderately silicified rhyolite has D = 0.23 - 0.26, clay-altered andesite has D = 0.3, strongly silicified rhyolite has D = 0.3 - 0.39, silicified andesite has D = 0.4, and silicified black shale has D = 0.43. The D values are related to the composition of the host rock, and appear to be controlled by its competency. Hydrothermal silicification results in increased competency and higher D values. Vein thickness frequency plots for different lithologies show that andesite at Ohui, and to a much lesser degree at Wharekirauponga, contains more thick veins than rhyolite and tuff. Gold grades show a strong to fair correlation with higher vein density in three drillholes at Wharekirauponga. The fractal dimension of veining is an important control on fluid flow and gold deposition within an active epithermal system. Vein systems with higher D have a greater density of veining which would have favoured increased fluid-rock interaction. Lower D vein systems have more clustered arrays. The increased fluid-rock interaction induced by a high D vein system, when combined with a reactive host rock, results in an environment which is highly favourable for gold deposition. This situation is seen in the andesite at Wharekirauponga.

Jan 1, 1994

(Published by courtesy of the Great Boulder Pty. Gold Mines, Limited)1. INTRODUCTIONThe faults described in the following paper are of a somewhat unusual character, and have presented a serious problem both in the development and mining of many ore-bodies on the Eastern side of the Boulder Belt. They are mentioned only briefly in the literature of the field, but were observed by, both Stillwell (1)t and Gustafson (2), and have been studied by other geologists and operators in various parts of the "Belt." Their effect on the lodes of Hamilton Section, Great Boulder Mine, was described by King (3) in 1937, and, during recent years, they have been encountered in most of the Great Boulder workings extending east of the Boulder Dyke.Although this paper is based mainly on a study of faulting and development problems on the Great Boulder mine, the general conclusions reached may be applicable to the solution of similar faulting problems in other mines throughout the field.In order to arrive at a clear understanding of the somewhat complicated relationships which the east-dipping faults bear to the lode systems it will be necessary to give a brief outline of the salient geological and structural features of the "Belt" before proceeding to describe the faults themselves.2. GENERAL GEOLOGYThe most recent descriptions of the geology of Kalgoorlie are those by Stillwell (1) and Gustafson (2) and the reader is referred to these as representing the basis of current knowledge of the geology of the field.The Boulder Belt is comprised of a central core of altered quartz dolerite which is flanked on either side by fine grained greenstones and lavas, usually referred to as the calc schists. Down the centre of the quartz dolerite mass there is an infolded bed of slate (2, p. 105) which has been intruded in lit par lit fashion by albite porphyry; the composite body of slate and porphyry so formed is known as the "Boulder Dyke." Numerous smaller porphyry dykes intrude both the quartz dolerite and the calc schist; these vary in composition and range from chloritised hornblende porphyrites to albite porphyries.The whole of the rock assemblage has been subjected tG a considerable amount of folding, faulting, shearing and fracturing, the greater part of which took place before the formation of the lodes.

Jan 1, 1941

By Douglas D. Reynolds

Intense vibration can be transmitted to the hands and arms of workers who use hand-held percussive or vibrating devices, tools, and work pieces. Continued habitual exposure to vibration directed to the hands can cause patterns of various symptoms associated with hand-arm vibration syndrome (HVAS). The International Organization for Standardization (ISO) first published ISO 5349 in 1986.3 This standard specified methods for measuring and evaluating vibration directed into the hands from hand-held vibrating devices, tools, and work pieces. The American National Standards Institute (ANSI) published ANSI S3.34 the same year.1 This standard was modeled after ISO 5349-1986 and specified methods for assessing exposure to hand-arm vibration. The Parliament of the European Union has issued the European Union Human Vibration Directive-2002/44/EC, which specifies vibration daily exposure action values (DEAV) of 2.5 m/s2 and daily exposure limit values (DELV) of 5.0 m/s2. These values have generally been accepted by medical experts, scientists, and engineers in governmental agencies, research institutions, and industry in the USA and other countries.2 When they are achieved, they will reduce the potential for the development of symptoms related to HAVS among workers exposed to hand-arm vibration. Significant improvements in measurement and analysis instrumentation, miniature and subminiature accelerometers, and medical diagnostic and assessment protocols have being introduced since 1986 when ANSI S3.34 was first published. In response to these improvements and the introduction of the EU Human Vibration Directive, ANSI Working Group S2.39 developed the revision to ANSI S3.34, which has now been published as ANSI S2.70-2006.2 Method ANSI S2.70 specifies the use of the hand-arm vibration measurement procedures outlined in ISO 5349, Parts 1 and 2.2,4,5 It requires the measurement of ISO frequency-weighted acceleration values in three mutually orthogonal axes of vibration. These values are then vectorially added to obtain the vibration total value, ahv: [ ] where ahwx, ahwy, and ahwz are the measured r.m.s. ISO frequency-weighted acceleration values in the x, y, and z directions, respectively. If multiple vibration exposure events are experienced during a work day, the overall vibration total value is obtained from: [ ] where ahvi is the vibration total value of the ith operation, Ti is time duration in hours of the ith operation, n is the total number of operations, and T is total time in hours associated with the n

Jan 6, 2006

By C. Nexhip, S. Sun

Critical to understanding and controlling phase mixing and separation in metallurgical processes, is the development of reliable experimental techniques to measure interfacial properties of melt systems. A series of experiments has been carried out to determine the effect of oxygen potential on the surface tension of liquid copper. Using an electromagnetic levitation method, a surface tension-oxygen chemical potential adsorption isotherm was produced for the Cu-O system at T 1400 K. The surface tension results show good agreement with the literature values e.g. within 3 per cent for pure copper. The Belton `ideal' adsorption isotherm was applied, and the following equation derived to best fit the experimental data ? =1251-77.5 1n41+ (2.79x105 . pO1/2)](mN.m Using this equation, the surface tension of copper was calculated/predicted as a function of oxygen potential at 1400 K, and showed excellent agreement with the experimental values. This confirmed that the adsorption behaviour of oxygen in liquid copper is ideal. The saturation coverage of oxygen on the surface of liquid copper was determined to be T0 = 6.65 µmoles.m,-2 corresponding to an area per adsorbed oxygen atom of ao ? 25 Å2. These values agree well with those cited in the literature, and our experimental data cover a very wide range of oxygen potential (10-5 to 10-17 atm). Keywords: levitation melting, surface tension, Cu-O alloys, oxygen potential, adsorption isotherm, surface coverage.

Jan 1, 2004

By B. R. Reddy, Y. V. Swamy, B. V. R. Murthy, H. S. Ray

The Ni oxide was reduced with a CO-CO2-N2 gas mixture and the effects were observed of process variables such as roasting temperature, time and CO partial pressure. Extraction of over 90% of Ni was achieved by ammoniacal leaching of ore that had been reduced for 1 h at 760 degrees C under a gas mixture with a volumetric CO/(CO+CO2) ratio of 0.57. The kinetic data were found to conform to the Ginstling-Brounstein equation. The applicability of this model was confirmed by evaluating the apparent activation energy with the use of independently constructed differential and integral plots, which showed an excellent match

Dec 1, 1995

By R. J. Leary

The technologies of contemporary iron and steel production and of direct reduction processes have been studied to identify and estimate the potential effects of direct reduction on mineral consumption and the problems raised thereby for the Nation's mineral supply to the iron and steel industry. Contemporary patterns of mineral raw material and energy consumption were determined from a study of the 1964-70 period. Major changes were observed in requirements for 14 of 20 principal commodities. Average practice in 1970 was compared to best practice to estimate the potential effects of implementing existing technology without recourse to direct reduction. Potential effects of direct reduction were estimated by comparing the mineral requirements for 95 percent reduced pellets with 1970 best practice technology for two cases of use: (I) complete substitution for hematite pellets in blast furnace burdens, and (11) partial substitution (40 percent) for scrap in electric furnace steel charges. Effects upon mineral supplies are described; important collateral factors are analyzed, and the principal metallurgical, industrial, and environmental consequences of implementing direct reduction are discussed. Major effects of direct reduction on the domestic steel industry and its mineral supply system are foreseeable. Direct reduction will (1) create a large demand for low-gangue iron ore; (2) relieve the dependence of the steel industry on coking coal, and (3) permit great increases in furnace productivity, both in blast furnace ironmaking and electric arc furnace steelmaking.

Jan 1, 1973

By Maxwell Gensamer

Jan 1, 1960

By James C. M. Li, J. Weertman

James C. M. Li (United States Steel Covp.)—The author has discovered a single analytical relation between the minimum, or the steady-state, creep rate and the applied stress confirming a statement made earlier by orn" that such relation should exist and should reduce to the power law at low-stress levels and to the exponential law at high-stress levels. However, as pointed out by the author, all the available theories of creep are not adequate to explain such a single relation. Inspired by such a statement, I have made a suggestion concerning a possible explanation13 of the new relation. This is based on the concept that the rate-controlling step is the cooperative motion of a short dislocation segment. The nature of and the reason for the cooperation have been discussed13 as well as the reason why the motion of the dislocation could be controlled by the mobility of the short segments. A natural short segment has been proposed13 to be the node formed at the intersection of two dislocations. The reason why the node may contain only a few atoms has not been discussed and the application of the proposed model to all the metals and alloys reported in this paper has not been attempted. It seems appropriate to include these in this discussion. The model was partly inspired by the substructure recently observed in the electron microscope on creep specimens. One such example is shown in Fig. 48 in Darken's Campbell Memorial Lecture. The substructure is a collection of dislocation intersections and irregular networks. No pileups have ever been observed in such specimens. A closer look at the network indicates that one set of dislocations is more straight than the other set sug- gesting that the other set may be mobile in the plane of the net. Since, for the network to survive the creep deformation, some dislocation reaction has to take place at the intersection of the two sets of dislocations, the mobile dislocation, even if it can slip in the plane of the net, will have difficulty in moving at these intersections. To illustrate, an idealized network is shown in Fig. 9. The thin lines represent the set of dislocations which are mobile and can slip in the plane of the net except at the nodes which are formed at the intersection of the two sets of dislocations. The other set of dislocations, represented by the thick lines, may have a Burgers vector which is not parallel to the plane of the net, and therefore may be immobile. Because of this nonparallel Burgers vector, the Burgers vector of the node, which is avectorial sum of the Burgers vectors of the two sets of dislocations, will also be nonparallel to the plane of the net. Therefore, the node cannot slip in the plane of the net and the set of mobile dislocations will have to drag the node along as indicated in Fig. 9. Now the problem is how many atoms could be involved in the nodes. Exact calculation of the length of the node is a difficult task especially since the length should be a steady-state length and not an equilibrium length. An estimate of the equilibrium length can be made by balancing the line tension of all the dislocations. The results would be a small fraction of the length 1 between nodes as shown in Fig. 10. In some instances such a calculation shows that the node will not be formed. But, since the calculation is only approximate, it does not rule out the possibility of the formation of a short node. In other instances the equilibrium length of the node is an appreciable fraction of I. Such a node would

Jan 1, 1963

By Axel Nitschke, Josef Arnold

Bochum’s Underground Lot 306 was characterised by a confined site and unique tunnelling methods. A highly mechanized construction concept was developed to achieve high progress rates. Bespoke backup systems for muck disposal and material supply were used. For the first time in suburban tunnelling, wet shotcrete, batched onsite, was applied. A three-storey underground car park had to be undermined with only1.5 m of overburden and had to be lifted several times by “compensation grout in.” An existing operational two-lane metro tunnel was enlarged without any major impact on train operation.

Jan 1, 2005

By M. Melashvili

"The SART Process was introduced to the gold mining industry in the late 1990s and several commercial plants were built over the following 10 to 15 years. Most of these plants have operated well and enjoyed plant performance in line with expectations based on the underlying chemistry of the process. However, two of the plants have consistently suffered from high reagent consumption and poor product quality, which has motivated a re-examination of the process fundamentals in the hope of understanding the reasons for poor performance when it occurs, and developing mitigating strategies.This testwork along with a review of the relevant literature has demonstrated that the amount of sulphide added to the primary reactor, relative to the copper concentration, is the most critical parameter dictating plant performance, with pH and residence time important secondary factors. The primary product in the SART Process is synthetic chalcocite, Cu2S, and the stoichiometric requirement for sulphide addition is therefore 0.5 moles per mole of copper in solution. Testwork shows that when sulphide is added in excess of this amount (generally with the express purpose of lowering residual copper in solution) it not only fails in its intended purpose, but can have a negative impact on the settling and filtration characteristics of the copper sulphide product, as well as the final product quality. Moreover, these problems are exacerbated the higher the selected operating pH in the pH 3 to 5 range, and the longer the residence time of slurry in this pH range.Freshly precipitated copper sulphide compounds are known to be very fine and slimy, with poor settling and filtration characteristics. It is therefore important to engineer the SART Process in such a way as to encourage growth in the average particle size of the precipitate crystals. This was the rationale for recycling a portion of the thickener underflow to the primary reactor when the SART Process was developed. The theory of crystallization predicts that compounds of extremely low solubility, such as chalcocite (Cu2S: KSP = 10-48), are likely to experience growth in crystal particle size through aggregation rather than heterogeneous nucleation. On this basis, the SART practice of recycling a portion of thickener underflow to the primary reactor in the commercial plants is unlikely to lead to crystal growth via nucleation, and could in fact promote breakage of the aggregates that formed in the thickener. Recycling thickener underflow could therefore have the opposite of the desired outcome. The result of testwork that is reported here lends credence to this theory and questions the wisdom of recycling in the SART process.There is evidence to suggest that operating the copper sulphide thickener in the SART Process without flocculants may be beneficial. Thickener underflow densities could improve from the current 10 - 15% range to 40 - 50%. This will not only reduce the flowrate of slurry to the filter, but the filtration properties of the slurry (filtration rate, filter cake moisture content) will also improve. This is one of the most critical engineering and design aspects of the SART Process, and the achievement of high solids density in the thickener underflow will not only improve the health and safety aspects of the operation, but also has the potential to lower capital costs and improve the quality of the final product."

Jan 1, 2016