Search Documents

By N. Mejia, Z. Xiao, D. Fragomini, A. R. Laplante

"The adaptation to platinum-group minerals (PGMs) of a protocol developed to predict gold recovery by gravity is presented. The protocol is based on three different data sets: ore characterization of gravity-recoverable minerals, description of their behaviour in cyclones and ball mills, and performance of recovery units. Examples of these are given. It is shown that the gravity-recoverable PGM (GRPGM) content can be very high. A case study at the Clarabelle mill is presented. Sperrylite is the dominant platinum mineral, and its partition curve is very similar to that of gold, despite its lower density. Its grinding kinetics is significantly higher than for gold. As a result, sperrylite accumulates in finer sizes than native gold, but can still be recovered by gravity or flash flotation. Recovery projections based on gravity recoverable data are presented.BACKGROUNDReliable modeling of gold recovery by gravity or flash flotation from a grinding circuit is based on the ability to describe the properties of gold-bearing particles in the ore, their behaviour in grinding and classification units and the performance of the equipment used for recovery, be it gravity or flotation. This is shown in Figure 1."

Jan 1, 2004

By H. Y. Sohn, R. D. Hagni, M. K. Kojo

Reflected light microscopy was used to study the mineralogy and textures of experimental reactions of a nickel concentrate with a slag. Such an interaction is an important step in the bath smelting of nickel sulfide concentrate. Concentrate particles in the central portions of the pellets are rounded and develop vermicular SO2 gas porosity. In the outer portions of the pellets, the original minerals of the concentrate, pyrrhotite, pentlandite, pyrite, chalcopyrite, and violarite have agglomerated and reacted to form two intergrown phases: 1) pyrrhotite solid solution (Fe0.75Ni0.18Cu0.01S) and pentlandite solid solution (Fe0.58Ni0.51Cu0.03S). Magnetite formed euhedral crystals, skeletal crystals, and dendrites at the margins of the matte. Prismatic ferrohypersthene and octahedral Mg-Al-Fe spinel crystals and dendrites are present in the glassy slag.

Jan 1, 2001

From a vast army of workers to gargantuan machines to immense volumes of material being moved, almost everything about the Port of Miami Tunnel (POMT) project points to tunneling on a massive scale. Almost.The star of this underground and underwater show is a Herrenknecht tunnel boring machine (TBM,) nicknamed ?Harriet? by a local Girl Scout troop. With her 12.9-m- (42.3-ft-) diameter cutting head and a total length of 139 m (457 ft), Harriet is boring through thousands of cubic meters of clay, dirt and sand, which is often mixed with coral and other materials, to create two 1,280-m (4,200-ft) traf?c tunnels connecting Watson Island and the Port of Miami (Dodge Island) beneath Biscayne Bay, FL. But in the midst of all of the oversized equipment and machines, a strong contender for the best supporting nod goes to a relatively tiny machine: the Brokk 400. Bouygues Civil Works Florida crews are using the compact, remote-controlled and electric-powered machine to excavate five cross passages between the twin traffic tunnels. The pas-sages will be essential to the safety of people who drive their vehicles under the bay every day.

Sep 1, 2013

By R. D. Leitch

"INTRODUCTION The character of mine drainage from areas of low pyritic coal3 and from average pyritic coal4 has been investigated, and to complete the study it seemed desirable to select a number of mines situated in high pyritic are as to determine whether or not this variation determines the relative acidity of drainage from mines operating in these characteristic coal fields. Advantage has been taken of some worked on in 1918 by Henry Leighton 5, then Professor of Economic Geology, University of Pittsburgh. He visited and has reported certain areas of high pyritic coal which we also visited during this study. Many of the mines were found to be idle in 1930 and a few were caved shut, apparently for several years. However, samples of the outflow from all mines visited were collected if possible and analyzed, and some general information was gathered for each mine. The map (fig. 1) shows the location of these areas as given by Leighton.ACKNOWLEDGMENTSThe information and assistance obtained from officials and others at the mines visited during this study are gratefully acknowledged, as well as the original suggestion from Dr. Glenn V. Brown of the State Sanitary Water Board who called attention to the desirability of such a study.Thanks are also especially due the coal laboratory of the United States Bureau of Mines, under the direction of H. M. Cooper, for the analyses and determination of sulphur forms in the large number of solid samples necessary to the work reported,This study has been carried on under the direction of R. R. Sayers, chief, health and safety branch, and W. P. Yant, supervising engineer, Pittsburgh Experiment Station, of the U. S. Bureau of Mines."

Jan 1, 1932

Chelopech Mining EAD and Best Available Techniques for Management of Waste Rock in Mining Activities

Sep 13, 2010

By David A. Hill, James R. Wait

This report contains a summary of analytical research on electromagnetic methods of nondestructive testing of cylindrical conductors. The principal objective is to provide theoretical insight to the operation of electromagnetic sensing of the internal properties of wire ropes that are used extensively in mine hoists. An overall review of past work is also included.

By L. Bergkvist

"History can provide many answers with regard to what today’s mining industry looks like and how we got to where we are today. It has always been self-evident that safety has to be given the highest priority and, throughout history, many mine workers have been injured, become ill and have even sacrificed their lives due to the way in which attempts have been made to extract the ore, often due to ignorance of the dangers being faced. In more recent stages of history we have seen certain dangers being compensated by means of higher wages and bonus systems in order to have the work carried out: one example of this is when the rock heating method was employed and the men who first went underground to put out the fires that were burning, were paid extra. The working conditions, to which these people were exposed, with smoke-filled closed spaces, shortened their life expectancy considerably and today we can ask ourselves whether it is really possible to value these efforts just by higher wages. In all mines worldwide the working environment has been improved and many mining companies have a zero vision where work-related injuries and accidents are concerned. Despite this, incidents of varying severity still occur and a continuous improvement process and ongoing training is necessary in order to constantly minimise risks and improve the working environment. Atlas Copco takes this into account in all its development and we not only review our own production units but also the needs of our customers, with regard to the machinery that is produced in order to cater for customers’ safety and production objectives. Underground mining probably started at the beginning of the 11th century when people with modest tools tried to extract ores that were present in the bedrock. Methodology has developed enormously over the years with a strong focus on increased efficiency and higher production. In different parts of the world the development is driven forwards at different rates depending on which controlling factors are regarded as necessary. Political decisions, profit goals and productivity have partly controlled the development, but the greatest factor where improvement is concerned still remains with increased understanding of the totality. During more than 1000 years since underground mining has been conducted, all individuals at all levels have constantly improved their expertise and the human element has been of great importance with regard to the development that has been made, even though technology often occupies a leading role when we describe the improvements made."

Jan 1, 2011

By B. P. Nutter, M. Lebourg, J. A. McAlister

The Multi-Flow Evaluator (MFE) is a new system of tools providing an original approach in drill-stem testing. It improves control during the testing operation and gives a more accurate evaluation of the fluid recovered while providing additional pressure information for reservoir analysis. The tools are operated entirely by up and down motion of the drill stem. This up and down operation provides a positive means of control and offers easily observed surface indications of tool operating position. An unlimited number of shut-in and flow periods may be taken with this tool while in the hole. The key to the success of the MFE system of tools is the safety seal packer. Until this development, operation of the tools by up and down motion only had proven unreliable. The success ratio of the MFE now exceeds the ratio achieved by conventional tools. The equipment includes a 2,750-cc chamber in which a representative sample of the flowing formation fluid is trapped at the end of the last flow period and brought to the surface under pressure. The sample can be evaluated at the wellsite or transferred under pressure for laboratory analysis. The sampling feature of the tool allows a sample to be obtained from a reservoir which has suffered minimum influence from production. A representative sample can be obtained for laboratory or empirical analysis by employing a testing technique to minimize drawdown during flow periods. Interpretation methods to take advantage of this additional information are presented and supported by actual field examples. INTRODUCTlON The Multi-Flow Evaluator (MFE) represents the most recent technical advance in formation testing. Drill-stem testing techniques prior to the MFE have been limited in three areas—(1) operaiion, (2) recovery analysis, and (3) reservoir analysis. The tools in the MFE system allow maximum control during testing and provide test information superior to that obtainable with conventional tools. The system of tools and the operation procedures are described to illustrate the mechanical advantages over present tools. Methods of pressure build-up and fluid

Jan 1, 1966

By C. S. Samis, D. P. Seraphim

In the presence of oxygen, galena is oxidized in an aqueous medium containing ammonium acetate in accordance with the following reaction: PbS + 1/2 0, + 2 NH~Ac -» PbAc, + So + 2 NH: + H2O. This reaction was studied in an autoclave under oxygen pressure by polarographic measurement of the concentration of lead ion in solution. Oxidation rate is parabolic. The reaction products are lead acetote and elemental sulfur, the latter forming a film on the galena crystal. Reaction kinetics appear to be determined by the diffusion of lead from the galena phase to the solution interface through the film of sulfur. RECENT applications of pressure leaching on a commercial scale have emphasized the metallurgical importance of the oxidation of sulfide minerals in aqueous media by oxygen under pressure. Research experience has shown that sulfates, thio-nates, and metal oxides are the products of humid oxidation of sulfide minerals in basic solutions. The metal oxides may be solubilized in the presence of a complexing ion.' Certain oxide minerals are also amenable to treatment by this type of oxidation.' Several kinetic studies have been made of reactions of this type. Stenhouse' found that pyrite (FeS?) oxidized in a sodium hydroxide solution to produce a soluble sulfate and an insoluble layer of iron oxide and that the kinetics of the reaction were controlled by oxygen diffusion through the oxide layer. Andersen' studied the oxidation of galena (PbS) in a sodium hydroxide solution under oxygen pressure. He found the lead and sulfur to be simultaneously oxidized to produce plum bite ions and sulfate ions, and that the reaction rate was proportional to the surface area of the galena. The kinetics appeared to be determined by the reaction of oxygen and water with the galena surface. In each of the above reactions the metal and sulfur components of the mineral were oxidized simultaneously and at equal rates. In the present study the alkaline medium has been replaced by an approximately neutral solution of ammonium acetate in which lead sulfate is soluble. Under the conditions no sulfate was produced and no sulfur could be detected in the solution. The metal component of the mineral was solubilized and the liberated sulfur remained as a film of elemental sulfur on the surface of the galena crystal. A kinetic study of this reaction was undertaken to establish the kinetic mechanism and to determine the rate controlling step. Specific reaction rates were determined by oxidizing crystals of galena of measured surface areas in an autoclave maintained at the desired temperature and oxygen pressure. The rate of reaction was followed by measuring the concentration of lead in the ammonium acetate solution with a cathode ray polarograph, employing stationary platinum electrodes incorporated in the autoclave. Experimental Materials and Equipment Crystals of galena obtained from Violamac Mines, Retallack, B. C., were selected for oxidation because

Jan 1, 1957

By J L. Lett, I Brunton, T Thornhill

Block cave fragmentation has been documented in the literature to have a significant impact on mine design, performance and downstream material handling and comminution processes. Due to this impact, emphasis is placed on the premining prediction and subsequent quantification of primary and secondary fragmentation distribution. This paper discusses the fragmentation prediction and assessment methodology employed at the Ridgeway Deeps (RWD) and Cadia East PC1 (CEPC1) cave operations.Prediction of drawpoint fragmentation was based on an iterative process, which relied on an improved quantitative and qualitative understanding of the rock mass as the mining project developed. In particular, the collection of scanline mapping data provided detailed information in relation to discontinuity trace length, persistence, termination characteristics and large-scale planarity. This allowed an improved estimation of select modelling input parameters, in particular ‘block forming’ discontinuity orientation and strength characteristics.With the start of production, drawpoint fragmentation assessment was undertaken using digital image processing (DIP) techniques for drawpoint extracted tonnes up to 61 000 t. This technique provided a consistent methodology to estimate changes in fragmentation as the block cave matured. It can generally be concluded that the measured mean fragmentation distribution became finer as drawpoint tonnage increased. The incremental change in fragmentation distribution (per 10 000 t) was not uniform, with step changes evident. In addition, for a given geotechnical domain and drawpoint tonnage, the measured fragmentation distribution range was found to vary significantly. A possible explanation for these observations may relate to the disturbed flow behaviour noted from marker experiments at RWD and CEPC1. These experiments indicate that material reporting at the drawpoints (for any given tonnage extracted) may be sourced from highly variable heights within the cave, with finer material preferentially reporting.A comparison between predicted and measured drawpoint fragmentation distribution indicated two general findingsa close match between predicted and measured distributions for material greater than 1.0 m3an under prediction of finer material passing 0.5 m3 relative to measured.The findings would indicate that the fragmentation modelling frameworks utilised at both RWD and CEPC1 were deficient at predicting the percentage of finer material for both primary and secondary fragmentation.The wide range in fragmentation distributions observed in the measured results, and the discrepancy between modelled and measured fines should be considered in the design of mine (material flow, extraction level geometry and draw rate), material handling (crushing and material movement), and downstream milling (comminution and metal recovery) processes for future block cave mines at the Cadia Valley Operations (CVO). With respect to fragmentation distribution range, a sensitivity analysis should be undertaken for these design considerations, with a range in predicted fragmentation distributions investigated (for example, lower quartile, mean and upper quartile). For future mines with similar geotechnical and mining conditions, the calibration of fines should be considered in any fragmentation distribution prediction to match those measured previously.CITATION:Brunton, I, Lett, J L and Thornhill, T, 2016. Fragmentation prediction and assessment at the Ridgeway Deeps and Cadia East cave operations, in Proceedings Seventh International Conference and Exhibition on Mass Mining (MassMin 2016), pp 151–160 (The Australasian Institute of Mining and Metallurgy: Melbourne).

May 9, 2016

By M. B. Boyer, C. H. Herty, F. A. Hartgen, G. L. Frear

"DEFINITION OF VISCOSITYViscosity has been defined as the resistance experienced by one portion of a liquid moving over another portion, such as one molecular plane moving over an adjacent plane. A measure of viscosity is the force required to keep one plane in motion at Unit velocity, the other plane being considered stationary. Figure 1 illustrates the general idea that the surface friction between two adjacent areas will be greater the larger the molecules. Therefore it may be assumed that viscosity will be increased by molecular association or agglomeration, and likewise dissociation rill decrease viscosity. According to Maxwell (1)6 the unit of viscosity, the poise, is the tangential force per square centimeter of area of either of two horizontal planes 1 centimeter apart, one of which is fixed while the other moves with unit velocity, the space between being filled with viscous liquid of unit viscosity.The need for a satisfactory apparatus and experimental procedure for the determination of viscosity of high-melting materials at temperatures corresponding to those in pig-iron and steel making has been recognized during the past few decades. The condition of the slag during the smelting and refining periods in the iron- and steel-making processes exerts an important influence on the metal product obtained, this slag condition really meaning the state of viscosity or, vice versa, fluidity, which the slag attains under certain conditions."

Jun 1, 1934

By D. Turk, P. Lefebvre, J. Sargent, K. Major

"Stillwater Mining Company recovers platinum and palladium from their Nye Montana operations. Mining first commenced at Nye in 1987 at a milling rate of 455 t/d (500 stpd). The plant was expanded in 1995 with the replacement of the ball mill and the addition of flotation capacity to increase mill throughput to 1820 t/d (2000 stpd). The applications for platinum and palladium have increased in recent years particularly in the automotive industry and the supply of the metals has remained fairly constant with prices for both platinum and palladium increasing and remaining very stable. In 1997 Stillwater initiated the expansion of their operations to increase production of platinum and palladium. Projects were initiated to expand production at the Nye operations, increase smelter and refining capacities at the Columbus facilities and to develop the extension of the orebody at the East Boulder site. This paper will discuss the changes at the Nye concentrator which were completed to increase production from 1820 t/d (2000 stpd) to 2730 t/d (3000 stpd) and eliminate some of the process bottlenecks.IntroductionStillwater Mining Company is mining the only significant primary source of PGMs outside of South Africa. The J-M Reef is an extensive mineralized zone containing PGMs that has been traced over a strike length of about 43 km (27 miles) in the southwest region of Montana. Stillwater controls the entire PGM mineralized zone.Johns-Manville geologists discovered palladium and platinum in the early 1970’s. In 1979, a Manville subsidiary entered into a partnership with Chevron, USA to develop the project. In 1 987, operations at the Stillwater property near Nye, Montana commenced."

Jan 1, 2000

By Andréa Lagacé

"The milling of sulphide ores inherently generates “thiosalts”, a series of partially oxidized sulphur oxyanions (S2O3 2-, S3O6 2-, S4O6 2- etc.), which make control of effluent treatment systems difficult because their natural oxidation produces acid with the potential to lower the pH within the treatment system and/or receiving environment.At the Xstrata Copper Kidd Metallurgical Site, thiosalt oxidation in the Tailings Management Area continues after buffer storage Ponds A and C where lime is added to precipitate metals and elevate the pH to within the required range for discharge. If the additional acidity generated during the continued process of thiosalt oxidation is not neutralized there is a risk that the final effluent will be non-compliant with regulated standards.To mitigate this risk, in 2007 a full scale pilot trial was conducted with hydrogen peroxide in order to force the complete oxidation of thiosalts prior to the lime treatment stations. This method of chemical oxidation had been shown to be effective at Xstrata Zinc’s Brunswick Mine. The pilot results indicated that a maximum of 50% of the stoichiometric quantity of hydrogen peroxide is required to reduce the thiosalt concentration in the No. 2 thickener overflow stream (feed to buffer storage Ponds A and C) to less than detection (10 mg/L); however, treatment was only effective in the presence of a 20 mg/L iron (Fe) catalyst. Since the hydrogen peroxide addition point at the No. 2 tailings thickener overflow typically contains <1 mg/L total Fe, a second reagent, ferric sulphate, was used to increase the total Fe concentration to the required level. This paper describes the set-up and results of the pilot trial which were used in the design of the full-scale plant.INTRODUCTIONTailings Management Area OverviewXstrata Copper’s Kidd Metallurgical Site, located in Timmins, Ontario, is an integrated metal refining facility producing copper, zinc, cadmium, indium, liquid sulfur dioxide, and sulfuric acid from third party custom feeds, and mineral concentrates recovered from its 7,800 tonne per day mill."

Jan 1, 2010

By Narendra N. Khatri, Michael Deutscher, Jozef F. Zurawski

Five sewer connections were required to divert Combined Sewer Overflow (CSO)from existing outfalls to an old underutilized large diameter brick sewer (the Easterly Interceptor). The connections were part of an on-going CSO control program by the Northeast Ohio Regional Sewer District (NEORSD) in the City of Cleveland, Ohio. Each connection was made by pipejacking and hand-mining a fiberglass-reinforced polymer (FRP) pipe through soft glaciolacustrine clays and silts. Jet grouting was utilized to stabilize the soils and fill the voids surrounding the interceptor prior to making the underground connection. Reinforced concrete liners were constructed inside the aging brick sewer to provide structural stability during ground improvements and connection work. The project also involved micro tunneling two approximately400 foot lengths of FRP pipe through soft ground using a slurry pressure balanced machine.

Jan 1, 2005

By T. P. R. de Jong, H. Harbeck

"This contribution reports recent mineral sorting experience and lists the opportunities of the newest generation of automatic sorters with a special focus on electromagnetic detection and Xray transmission. With the available sensor resolutions and actuator systems that are available today sorting of minerals down to 2 mm size and up to 250 mm lump size is technically feasible. In industry, ore as well as coal is still hand-picked and current automation potential only utilised in a handful of specific applications. Recently a number of technological developments took place that may increase sorting potential, save mill costs and lead to lower overall tonnage costs.They are:??High speed and high resolution sorting systems based on X-ray transmission imaging??High speed on-line image processing and further development of advanced data processing??Modular multi-sensor automatic sorting units (using one or more sensors such as colour, electromagnetic and X-ray in a single device)??Sensitive electromagnetic sensors for sorting based on differences in magnetic susceptibility and electric conductivityThe above enables compact dry separation, insensitive for moisture fluctuations and at high separation sharpness, as well as “electronic” high gradient magnetic separation for larger particles. Besides conventional X-ray transmission (XRT), dual-energy XRT (DE-XRT) is now being redesigned to meet the requirements of the mining industry. The mentioned technologies provide automatic dry separation of minerals and rock as alternative for conventional processing."

Jan 1, 2005

By Rose W. J

Geophysical prospecting was first recognized in Australia as an important aid to mining enterprises in 1927, when, on the suggestion of Sir Herbert Gepp, the &quot;Imperial Geophysical Experimental Survey&quot; was organized. The function of this survey was to carry out investigations on the applicability and efficacy of the most important geophysical methods under Australian conditions. The survey was formed under the joint auspices of the British Empire Marketing Board and Australian Commonwealth Government, and the investigations were carried out in Australia under the leadership of Mr. A. Broughton Edge, B.Sc., A.R.S.M., F.G.S., M.Inst.M.M., during the years 1928 and 1929. A description of the work performed, the apparatus used, the principles of the methods, and the results and conclusions of the survey are included in a report published by the Cambridge University Press in 1931.In 1932 Sir Herbert Gepp became interested in some gold mines in Eastern Victoria, in the Cassilis and Bethanga districts, and realizing that the lodes in these areas would be favorabIe for a geophysical examination, he arranged for a series of tests to be carried out. This work was performed by Mr. E. L. Blazey, B.E.E., a former member of the staff of the Imperial Geophysical Experimental, Survey, under the general direction of, and in consultation with, Mr. W. J. Rose, B.M.E.The work was commenced in October, 1932, at Cassilis, Vie., continued at Talgarno, Vie., .and concluded at...

Jan 1, 1933

By S Vianna, K Caldwell, R Silva, T Olson, D Bell

Studies to enhance the FLSmidth forced air machine through changes in the design, showed important improvements through changes in the number of Rotor blades, stator designs, rotor air distribution and other designs features. The main outcomes from this study highlighted that increasing the number of rotor blades increased the selectivity while decreasing power requirements while stators with slots have better results in terms of recovery, flotation efficiency and selectivity than without slots.Also the rotor air distribution makes a difference in the design when comparing different distribution criteria whilst keeping all features of shape and number of blades etc constant. At the most efficient rotor air distribution design, copper recovery reached one of the highest values while keeping good selectivity. The possible explanation for that is the increase in the pump capacity indicated by the increase in the specific power.It was seen as significant that reduction in specific power for some designs tested compared with the then current standard design had an average of 30% less power consumption.As a result, FLSmidth was able to identify new designs that decreased power and improved recovery, as well as, identifying some designs that specifically improved fine and/or coarse particle recovery.CITATION:Silva, R, Caldwell, K, Bell, D, Olson, T and Vianna, S, 2013. FLSmidth flotation developments - Dorr Oliver machines, in Proceedings MetPlant 2013 , pp 413-430 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 15, 2013

By Thomas Novak, Sampurna Arya

Controlling dust generation and keeping it below permissible limits to meet federal dust standards at the working face of a roomand-pillar coal mine is a challenge for a mine operator. With the recent changes in federal dust regulations requiring lower worker exposure, maintaining compliance has become increasingly difficult. The current most effective practice of dust control at a continuous miner face in an underground mine is the use of a flooded-bed scrubber. A study carried out by the National Institute for Occupational Safety and Health (NIOSH) indicated that a flooded-bed scrubber could achieve cleaning efficiencies between 58 and 90%. But, the operation of such a system is maintenance intensive. The flooded-bed scrubber screen becomes clogged with dust particles and requires frequent cleaning to maintain performance. However, the dust control issue is not solely a mining industry problem. Other industries face similar issues. The University of Kentucky collaborated with Toyota Motor Manufacturing on the development of a novel wet scrubber, called the vortecone scrubber, for capturing oversprayed paint particles in automotive paint booths. The vortecone scrubber achieved a cleaning efficiency of 99% and required minimal maintenance. This study aims to assess the ability of this vortecone scrubber to capture respirable dust in an underground coal mine. The paper presents the results of computational fluid dynamics (CFD) modeling of this vortecone scrubber. It discusses the effects of air quantity and dust particle size on the performance of the scrubber.

By Randy Vranes

Sophisticated information technology, quality management systems and a commitment to the environment has allowed Solutia’s Soda Springs phosphate mine and elemental phosphorus plant to reduce operating costs, improve product quality and contribute to the sustainable development of Idaho’s industrial minerals. Soda Springs background The Soda Springs phosphate mine and plant is a joint venture between Monsanto and its operating partner, Solutia Inc. The operation is based in southeastern Idaho in the Western Phosphate Field. The plant is located about 275 km (170 miles) northeast of Salt Lake City, UT and about 4 km (2.5 miles) north of the town of Soda Springs.

Jan 1, 1999

The relationship between neural network (NN) performance and the parameters that form the architecture of the neural network is complicated. This paper studies an issue that is still controversial—the effect of the number of neurons and number of hidden layers on NN performance. This paper tests if the above parameters affect network performance when the NN is trained according to the quick stop strategy. It was determined that the number of neurons did not have any effect on the NN performance for a single layer or a double layer network. It was also found that the training error and calibration error could be poor indicators of the prediction correlation coefficient.

Jan 1, 2003