Search Documents

By Mike Johnson, Niels Verbaan, David Dreisinger

The Search Minerals Foxtrot project in Labrador represents a significant Canadian resource of rare earth elements rich in the sought after heavy rare earth elements (19.8% HREE/TREE). A preliminary economic assessment of the project indicated that the project is technically feasible and economically attractive. A variety of metallurgical test programs have been undertaken which have led to significant streamlining and simplification of the process flowsheet. The improved flowsheet should allow Search to reach production earlier and will use conventional processing equipment. The results of these test programs are presented and discussed in this paper.

Jan 1, 2014

By G. F. Carr

"IntroductionFLUORSPAR has been found at several •localities in Newfoundland, but all the commercial deposits discovered to date are in the vicinity of St. Lawrence, a town on the southeast coast. Most of the veins are within six miles of St. Lawrence harbour, and between sea-level and an altitude of 400 feet. The St. Lawrence Corporation of Newfoundland, Limited, and Newfoundland Fluorspar, Limited, a subsidiary of the Aluminum Company of Canada. Limited. are the two producers.With the arrival of 1953, the St. Lawrence area entered upon its twenty-second •consecutive year as a producer of fluorspar. Production has increased more or less steadily during this period, as is shown by the fact that approximately one-half of the total shipments to date have been made during the past six years. Between four and five hundred men are employed by the industry, and more than $1,250,000 was paid out in wages and salaries during 1952. Shipments during that year reached an all-time high of 81,283 tons valued at $2,484,943, compared with 67,925 tons valued at $1,966,477 in 1951 (1).As a result of an agreement made in July, 1952, with the United States Defence Materials Procurement Agency, production of the St. Lawrence Corporation, Limited, and St. Lawrence Fluorspar, Incorporated, an affiliated company, will •be increased by .approximately 50,000 tons of acid-grade fluorspar per year. The contract extends over a period of four years, or until 150,000 tons of acid-grade concentrates have been produced."

Jan 1, 1954

By F. D. Ball, D. E. Gemmell

"Coal mining in New Brunswick dates from the year 1639, when, as today, coal was produced from the area surrounding Grand Lake near the geographic centre of the province. Current production by N.B. Coal Ltd. is approximately 453,000 tonnes per year and comes from a single, thin Pennsylvanian age seam. The history of development and understanding of the geological setting of New Brunswick coals is discussed, beginning with the work of A. Gesner in 1839. The lithostratigraphic sub-divisions of the Pennsylvanian strata proposed by various workers are summarized and the present hypothesis that only biostratigraphic subdivision is possible is detailed. The Pictou Group, which includes the New Brunswick coals, consists of repetitive sequences of sediments consistent with flu vial and upper delta plain sedimentation. The development of coal mining in the province has progressed from horse-powered stream-bank stripping operations prior to the late 1800s through to the present use of open stripmining techniques utilizing eight walking draglines of various sizes. During the intervening years the introduction and eventual phasing out of underground mining occurred with the near termination of the coal mining industry in the province prior to 1973. The coal seam worked by N.B. Coal Ltd. averages 46 cm in thickness, has never been found to be more than 76 cm thick and has both high ash (12.4 -31.1%) and sulphur (4.9-7.8%) content. The wide variety of coal beneficiation studies and tests completed on New Brunswick coals have lead to the conclusion that a washed New Brunswick coal product could be expected with 10 to 12% ash, 5 to 6% sulphur, and 30,225 to 31,388 1/gm. Recoverable coal reserves, including those in the area licenced to N.B. Coal Ltd. and those licenced, although not yet in production, to both the St. Lawrence Cement Co. Ltd. and Provincial Holdings Ltd. at Lake Stream and Beersvil/e, total 60 million tonnes. At present, coal in New Brunswick is removed from prospecting and staking and all existing and future coal exploration and development will be administered by agreement under the Mining Act. Based on the results of a recently completed drilling project undertaken by the Department of Natural Resources, seventeen coal exploration blocks which carry minimum work commitments will be offered to the New Brunswick Electric Power Commission on a sixty-day first right ofrefusal basis. Those blocks not taken up by the Power Commission will be offered under the same commitments to the private sector."

Jan 1, 1985

By J. R. Kadambi, E. Kucukal, J. Furlan

"Hydrocyclone separators are used in various industrial applications, including mining milling and tailings systems, in order to separate or classify solid particles and liquid droplets from multiphase systems. The flow within a hydrocyclone commonly used in the mineral processing industry was investigated numerically and experimentally using Particle Image Velocimetry (PIV) and refractive index matching techniques. The computational simulations were performed using a commercially available CFD package, STAR CCM+. Two turbulence models were used in the numerical calculations: the k-? model and the Reynolds Stress Model (RSM). In both cases, the first order discretization scheme was unable to resolve the flow characteristics accurately due to the high levels of numerical dissipation. PIV measurements were taken at various spatial locations along the cyclone. Both the numerical simulations and PIV data demonstrate that tangential velocities tend to increase when going radially inward from the cyclone wall towards the vortex finder, and reach a maximum value before they begin to decline rapidly and eventually become zero at the wall. The results also demonstrate the utility of non-axisymmetric approaches in cyclone flow simulations for a greater understanding of the fluid mechanics. The computational results were validated globally using pressure and flow rate readings at the boundaries and locally by comparison to the PIV velocity vector maps and profiles. The comparisons demonstrated good agreement between the two models.IntroductionHydrocyclones are devices used to separate and classify solid particles or liquid droplets from liquid-solid or liquid-liquid systems. Versatile in application, they have been widely used in many industrial applications including the food, oil, power generation, chemical and mineral processing industries since 1891, when the first hydrocyclone was introduced by Bretney (Park, 2003). Hydrocyclones have a relatively simple design with no moving parts, which leads to an efficient separation process."

Jan 1, 2016

By C. Díaz, M. Barati

Prof. Torstein Utigard's untimely death, in the spring of 2012, put an abrupt end to a meritorious career in the field of high temperature chemical process metallurgy. As an academic in the Department of Materials Science and Engineering of the University of Toronto, Torstein made solid contributions to the advancement of knowledge pertaining to production and refining of both nonferrous and light metals, and to the development of inquisitive and innovative young people in extractive metallurgy. In this profile, the authors highlight aspects of Torstein's rich personality and of his contributions to his field of expertise.

Jan 1, 2015

By Greg Yuriy

Over the past two decades, Vale Canada has used discrete-event simulation in a key role to support many underground mining studies. Traditionally, general purpose simulation software was developed for manufacturing and service applications. Not until recently have Mine Designers and Simulation Specialists been able to model the dynamics of underground mining scenarios without extensive manipulation of the inherent tools and entities contained within these simulation environments. This paper discusses Vale Canada?s past experiences with a variety of simulation software, describing model development methods and highlighting some of the lessons learned. The paper concludes with a review of recently considered software and provides a framework of desired attributes/functions for modeling mining scenarios.

Aug 1, 2013

By J. Abrahams, A. Nesbitt

Over the past half-century, researchers have commented on the complexities of ion-exchange processes, but have been forced to use gross assumptions to simplify the associated mathematics sufficiently for mathematical models to be of practical use, given the available computational power. This paper reviews the development of increasingly complex models as computational power has increased over the past half-century, from the direct application of Fick's Law in 1947, through to today's models requiring finiteelement methods. The assumptions required for and the associated limitations of the various models are presented and the associated computational complexities are discussed. A case study is presented in which a continuous loss of capacity in an acid resin was observed over about sixty cycles of loading and elution. Resin poisoning does not explain this observation. This paper presents the capacity loss data, speculates on possible causes of the phenomenon, discusses why conventional kinetic and equilibrium mechanisms cannot explain it and concludes that the need to develop a more rigorous model is apparent. The significance of this observation is that standard tests done for a new resin application could significantly overestimate the commercial performance of the resin concerned, highlighting the importance of computer simulation in this field of process engineering.

Jan 1, 2005

By S. Borle

The microstructural features produced in friction stir welded AZ31, AZ80 and AZ91 alloys are examined when different welding parameters are applied. The microstructures primarily consist of a-Mg grains with a-Mg + Mg17A112 colonies, however melted films of Mg17A112 are produced in AZ91 alloy when high rotation speeds are applied. When the rotation speeds are slow, or the travel speeds are fast the Mg17A112 were incorporated into the stir zone, however they do not exhibit an elongated lamellar morphology associated with melted films. In the case of AZ80, the extruded base material contains a smaller fraction of Mg17A112 and few of these phases are observed in the stir zones. The results are explained based on the dissolution rate and melting temperatures of Mg17A112 controlling the overall morophology and fraction of phases in the stir zone.

Jan 1, 2011

By Terry Keough

Contents for discussion ?Current liner technology ?Innovation of magnetic products ?Development of magnetic wear liners ?Installation and change out method ?Uses and capabilities ?Internal wear liners ?Summary Vulcanized rubber ?Many of you are familiar with using vulcanized rubber to provide wear protection on mining equipment. For years, this was the standard for most mining equipment, and mines spent a lot of money fabricating and lining steel equipment.

May 1, 2010

By J. L. Thornton

Introduction The importance of coal in our economy is well established, but it may not .be so generally realized how important a part it plays in our daily life. Select a product - any product - and, if it is made in America, bituminous coal has touched it first. Everywhere we will find coal, derivatives of coal, or functions made possible by coal. Apart from the well known major uses of coal for domestic purposes, coking, utilities, railways, and manufacture of steel and cement, there are over 200,000 products derived from coal, from lipstick colours to nylon products and perfumes. An item of interest is the recent action of the Canadian Government in voting $90,000 for development of coal deposits in Yukon Territory to make coal available to the settlers in that area. The value of this move, in case of some emergency, is very evident. Coal in its several stages of formation as peat, lignite, sub-bituminous, bituminous, semi-bituminous, semi-anthracite, and anthracite, has been used for centuries. Coal, as such, was not very well known to the ancients. First reference to "magic stone', or coal, was in 371 BC by Theophrastus, a pupil of Aristotle. This substance was said to burn like wood coal or charcoal and was used by the smiths. The first record we have in modern times was in 852 AD, when sea coal was gathered by the cowled monks of Durham, in Britain, and even at that time a major problem was transportation. Our first record of coal in this continent was near Richmond, Virginia, in 1700. Commercial mining was not under way, however, until 1750. The anthracite fields (presumably in Pennsylvania) were opened in 1793. Even so, there was no extensive mining until about the year 1820. The famous Pocahontas coalfields were opened in 1883. The earliest record of coal mining in Canada is about 1746.

Jan 1, 1950

By A. C. Wilson

"In 1991, local prospector C. “Mickey” Clement recovered at least three alluvial diamonds from the Michipicoten River. Two of these stones were sent to the Department of Mineralogy, Royal Ontario Museum, where they were identified as industrial-grade diamonds with weights of 1.05 and 1.13 carats. Both stones were described as frosted and graphiteinclusion riddled.In 1995, prospector Sandor Surmacz and geologist Marcelle Hauseux discovered diamonds in bedrock approximately 20 km north of the Michipicoten River. An 18.1 kg bulk sample of a xenolith-rich lamprophyre, taken from an outcrop at the side of the Trans Canada Highway north of Wawa, yielded 1 macro diamond and 5 micro diamonds. All but one was gem quality (Sage, 2000).Since then, more than 36 occurrences of diamonds in bedrock have been reported in an area approximately 30 km2 in size, centred 20 km north of the town of Wawa. The occurrences are hosted within a sequence of unusual, Archean-aged, heterolithic breccias previously considered to have no economic significance. Ten kimberlite occurrences also have been confirmed in the Michipicoten greenstone belt in the last five years. Two of these kimberlites are diamondiferous.To date, expenditures on property acquisition and diamond exploration in the Wawa are estimated at CDN$120 M.This paper summarizes the geological observations made on these occurrences and describes the relationship between the occurrences and the regional geology."

Jan 1, 2004

By Jack Zhang, Chuck Edwards

"The versatility and specificity of rare earth elements (REEs) have led to their use in an ever increasing variety of applications in new technologies. Consequently, demand for REEs has increased significantly. However, to separate REEs safely and effectively is a complex and expensive process. While China has consistently invested in R&D of REE processing technologies, there have been only sporadic activities in REE mineral processing in North America in the past 20 years. This paper presents a review of REE mineral processing technology, providing an update on current capabilities in REE mineral processing. This paper reviews REE extraction from the major commercially valuable REE-bearing minerals such as bastnaesite, monazite, and xenotime, and also from the ion-absorption type of REE deposits. Currently, there are four major REE mining areas in China, one in the United Sates, and one in Australia. A review of the REE extraction processes in these major REE mining areas closes out the paper.INTRODUCTION The rare earth elements (REEs) include the 15 lanthanide elements, plus yttrium and scandium. The lanthanide elements are normally categorized in two groups: the light rare earth elements (LREEs), i.e. lanthanum through europium (atomic number (Z) = 57 through 63); and the heavy rare earth elements (HREEs), i.e. gadolinium through lutetium (Z = 64 through 71). Yttrium is usually grouped with the HREEs due to similarity in physical and chemical properties. Table 1 lists the REEs, atomic numbers, and abundances in earth’s crust (Taylor and McCleannan, 1985). Promethium is the rarest due to lack of stable isotopes. REEs with low atomic numbers are more abundant than those with high atomic numbers. More interesting is that the REEs with even atomic numbers are much more abundant than adjacent lanthanide elements with odd atomic numbers as shown in Table 1."

Jan 1, 2012

By C. Chapman, M. G. Reading, M. P. Amsden

Ecstall Mining Limited, a subsidiary of Texasgulf Incorporated, operates the Kidd Creek mine, a 10,000- tpd base metal producer, near Timmins, Ontario. Two different ore types are mined and milled copper-zinc and silver-lead-zinc. At the highly automated Ecstall concentrator, the ores are processed by selective flotation in three individual circuits. One of the most important metallurgical tools for process control is a sample system incorporating continuous onstream X-ray analysis. Recently this system was further automated by adding computer control of reagents for two copper-zinc flotation circuits. The computer programs, utilizing optimum seeking methods and smelter contract terms, have increased copper recovery 0.8% and zinc recovery 1.6%, and reduced reagent costs by 12%. Although not anticipated, copper and zinc concentrate grades were improved modestly as well.

Jan 1, 1973

By C. Jay Hodgson, P. J. MacGeehan

"The gold ore zones of the Campbell and Dickenson mines occur within a km-wide, 2 km-long zone of highly altered and anomalously fissile and deformed volcanic and subvolcanic intrusive rocks, cut by a series of mainly post-deformation mafic and quartz-feldspar porphyry felsic dykes.Three main periods of gold ore-forming hydrothermal activity are recognized. In the first, the mine rocks were extensively altered and gold and arsenic enriched; syn-volcanic auriferous veinlet and low-grade chemical sedimentary ores were deposited in what was a major discharge zone of a submarine hydrothermal system. The chemical sedimentary ores are closely associated spatially and temporally with domal rhyolitic extrusions. Because of its mechanical and chemical properties, this synvolcanic alteration zone became the locus of later deformation, and the site of the generation of auriferous hydrothermal solutions, perhaps by pressure solution processes at depth. These solutions deposited, synchronously with deformation and metamorphism, the second and most economically important type of ore zones in the mine: large foliation-parallel and foliation-oblique banded carbonate- 'chert' Rssure-filling veins.In the third and final mineralizing event, high-grade, pipeshaped zones of siliceous replacement-type mineralization, and quartz-arsenopyrite-gold veins, were superimposed on the previous mineralization types after most of the deformation, but before the emplacement of the late dykes. The source of heat and also possibly the fluids for the third mineralizing period may have been the dyke-forming igneous activity. The textures and mineralogy of the banded carbonate- 'chert' veins suggest a near-surface (<3 km) environment of deposition, but their relationship to the fabric of their host rocks indicates they were emplaced in an environment, presumably at considerable depth (5-10 km?), in which the rocks were undergoing intense penetrative deformation. This anomaly is unresolved."

Jan 1, 1982

By Matthew A. Perras

With growing populations, infrastructure projects are utilizing underground space to expand transportation networks, water supply and sewer disposal systems and hydropower facilities. These new projects require large diameter tunnels to meet the growing demands. As the tunnel diameter increases, not all rock mass properties are scalable and precedent experience becomes less reliable. One such non-scaleable property is lamination thickness. Numerical modelling was conducted to determine the influence of lamination thicknesses on a 16 m diameter tunnel. Five different areas of excavation response were found to exist for laminations ranging between 0.16 to 16 metres in thickness. For this tunnel diameter, critical lamination thicknesses of 6 metre and 1 metre were found to exist. Above 6 metres the excavation response was similar to isotropic models. Below 6 metres the stresses and yield are channeled by the lamination boundaries, increasing plastic deformation. Below 1m multiple lamination units become unstable and the displacements increase significantly and the shape and extent of the yield zone size changes dramatically. The yield zone is controlled by bed deflections and bed parallel shear. As the lamination thickness decreases tensile failure first begins in the haunch area and progressively extends above the crown until a self-limiting plastic yield zone shape is reach at lamination thicknesses below 0.4 metres. Conclusions drawn from observations at the Niagara Tunnel Project are those of the authors.

May 1, 2009

By A. Warczok, T. A. Utigard

"Slags from smelting of copper concentrates contains from 1 to 15% of copper. Copper is present in the form of copper or copper matte inclusions of size from 2 to 1000 micro, and as dissolved cuprous oxide. Removal of the copper requires the slag reduction, primarily reduction of magnetite and co-reduction of cuprous oxide. Thus, the kinetics of the slag reduction play an important role in copper recovery. Instantaneously, at the contact of carbon and a slag the gaseous film is formed. The magnetite and cuprous oxide reduction is going through a gaseous film, separating the slag and carbon of a coke or electrodes. The thickness of a gaseous film grows to a critical value and gas bubbles, containing CO and CO2, are created and rise up in the slag. Results of measurements of kinetic of the slag reduction with graphite showed high value of activation energy, close to the value of activation energy of Boudouard reaction. Strong temperature dependence of the reaction rate points out the thermal instability of the reacting system. It was found that the slowest process in the reduction stages determines the loop of CO and CO2 transfer and magnetite diffusion to the slag/gas film boundary. Strongly endothermic Boudouard reaction seems to cause thermal instability on the reaction surface. So, heat transfer to the reaction surface is very important. Due to low thermal conductivity of a slag the convection dominates in the heat transfer. So, any slag stirring accelerates the rate of reduction and coalescence of metallic copper or copper matte inclusions."

Jan 1, 2015

By Randolph R. Dreher, Michael L. Schreder

"In recent years, FFE Minerals has designed, installed and commissioned several large grinding mill circuits. This paper discusses the issues related to the erection and commissioning of large grinding mill circuits. The three circuits discussed are Collahuasi in Chile, Minera Los Pelambres in Chile, and Century Zinc in Australia. These circuits include large SAG and ball mills with both gear and gearless drive arrangements, as well as, both trunnion and shell supported mills. Aspects of these features are discussed with respect to their impact on installation and commissioning. While the normal challenges of mill erection and start-up were experienced, there were also a few unexpected challenges encountered partly due to the large scale of the equipment. An overview of the equipment and design ratings of these circuits is also provided.COLLAHUASIOverviewThe Collahuasi mills were started-up in November of 1998. The site is located in the Andes Mountains at an elevation of 4300 meters. The grinding circuit includes two lines featuring one SAG mill and one ball mill for each line. The circuit features a recycle conveyor around the SAG mills with provision for adding a pebble crusher in the future. The design tonnage of the grinding plant is 60,000 MTPD with an ore BWI of 10.5 kWh/MT.The SAG mills are 32’ x 15’ (1 3’ EGL) with 10,800 HP drives. The drive arrangements are dual pinion through pneumatic clutches. The mills are driven by low speed synchronous motors with CCV variable speed drives. The SAG mills are trunnion supported and use hydrostatic lubrication. Feed to the SAG mills is via a rock-box type feed chute on a cart. The mills discharge on to trommel screens with a bucket wheel to remove the oversize.The ball mills are 22’ x 36.5’ (36’ EGL) with 12,500 HP drives. The drive arrangements are dual pinion through pneumatic clutches. The fixed speed mills are driven by low speed synchronous motors with GE Quadramatic load sharing drives. The ball mills are shell-supported and feature multi-pad bearings with hydrodynamic lubrication. The mills are fed via a rubber lined sump box with feed spout on a cart. The ball mills also discharge on to trommel screens. Twenty-six inch (26”) diameter hydrocyclones are in closed circuit with the ball mills."

Jan 1, 2001

By B. Roy

The Somita plant processes low sulphide gold ore in Burkina Faso with a CIL process. After experiencing recurrent low gold recoveries, Somita undertook an investigation on the source of the problem. COREM conducted a mineralogical study on different pit samples. A series of tests was also conducted on a composite sample from the plant which shows a content of 3.4 g/t in gold and 0.48 % in graphite. A reference test in laboratory resulted in a recovery of gold of 75%, similar to plant recoveries for the evaluated ore. In these tests, even if cyanide is not deficient, it was diverted mainly into iron cyanide complex, copper cyanide complexes and by graphite adsorption. Causes of low recovery were also identified and quantified. Solutions to increase gold recovery were suggested such as increasing gold dissolution kinetics by O2 or by lead nitrate addition, change in activated carbon addition point location to the 1st cyanidation tank and finer grinding to increase liberation. Increase in the pH could also contribute to reduce cyanidation consumption. Following these recommendations, Somita changed carbon addition from tank 2 to tank 1 in the CIL circuit. This change led to an increase of more than 5% in gold recovery. A system for the addition of lead nitrate is also under implementation at the plant and benefit will be evaluated in 2013. Finally, pit management for cyanidation will be done based on mineralogy instead of gold content in the future.

Jan 1, 2014

By Vicki Tschirhart, Bill Morris, McMaster University

A new method for mapping faults within basement rocks underlying the Thelon Formation and glacial overburden was developed and tested in the Aberdeen Sub-basin. This method utilizes newly acquired aeromagnetic data, the Blakely algorithm for defining magnetic source edges, a calculated dip-direction map, a digital elevation model (DEM) derived from the Canadian Digital Elevation Database, and the positions of identified, inferred and newly mapped faults that are within and adjacent to this sub-basin. Combining these data revealed three age groups of faults. One group is seen only on the DEM; these faults are young brittle structures that have no effect on the sub-Thelon basement unconformity, and are not visible on the dip direction map. A second group is evident on the dip direction map but not the DEM; these faults are old basement structures that did not propagate upward through the Thelon Formation. The third group is expressed on both the DEM and dip direction map; these faults are also old structures that propagated upward during and/or after deposition of the Thelon Formation. The latter group of faults may be of greatest interest for uranium exploration, with reactivation increasing their potential to serve as conduits to transport uranium-rich fluids and focus deposition at or near the unconformity surface. This methodology has promise for comprehensive mapping of basin faults and tracking the tectonic development of the whole Thelon Basin through time.

Jan 1, 2013

By Dave Bulled

"This paper describes the design of an SABC grinding circuit using CEET technology applied to grindability data produced from testing 100 samples and distributed across the blocks of a mine plan using a geostatistical approach. The data distribution required the identification of the location of the sample points and a geological plan of the 130 million ton Ruby Creek orebody together with a mine plan of about 11000 blocks to be mined during the proposed 18 year life of the mine. The mill design was then made with consideration for the estimated hardness of each block in terms of both SAG and Ball milling with allowance for annual periods when the ore is harder to grind. Operating specifications were selected to achieve an optimum of capital and operating costs through use of the CEET program. The geostatistics assigned the statistical error that arises from having a limited amount of sample testing data to the hardness estimate of each block of ore in the mine plan. The final design could then include a quantification of risk and determination of safety factors based on the accuracy of block hardness estimates and CEET model fitting.INTRODUCTIONThe traditional approach to plant design involves the extensive testing of a single large composite sample or a small number of composite samples that are reputed to represent the ore body. It is accepted that the tests used can accurately measure in the laboratory the grindability of the sample by a process representing that to be used in the plant. The size of equipment required to achieve a specified throughput and product size is then calculated from one of a variety of models that have been developed by metallurgists over years of trial and error. Since these tests and models are tried and tested, they are accepted as reasonably precise; however during the operation of the resulting plant the design is sometimes found to be inadequate. It is then suspected that the flaw in the design process lies in the samples not being sufficiently representative of the ore body, since using only a single or small number of composite samples does not recognize the variability of the hardness of the ore, nor does it allow for the lack of precision in the hardness value used in the design.This paper outlines a design procedure suited to an orebody that is described by a geostatistical analysis of a reasonably large number of small samples of drill core. The analysis required the identification of the location of the sample points and a geological plan of the orebody, together with a mine plan of the blocks to be mined during the proposed life of the mine. The mill design was then made using the estimated hardness of each individual block with an allowance for annual fluctuations in the grindability of the ore. The geostatistical analysis assigned a statistical error to the hardness estimates of blocks and of each production year such that the final design included risk estimates and safety factors based on the possible errors that arise from having a limited amount of sample data (Kosick et al. 2002)."

Jan 1, 2007