Search Documents

By R. Tipman, T. C. Burnett

"A general computer program has been written to calculate material balances using multiple regression techniques. The program can handle metal assays or size distribution data obtained in multiple surveys of the flotation circuit. The ""best fit"" weight split is calculated between the input and product streams by minimizing the sum of squares of the residual errors using all of the available data. The program includes an assay adjustment feature in which each assay is recalculated to fit the ""best fit"" mass balance. Thus assay or sampling errors can be easily identified. Also the assays can be statistically weighted so the most precise measurements are given the most influence on the material balance. The computer output reports the mass balance, the actual and corrected assays, the percent metal recoveries and the significance and confidence levels for the weight splits.The program enables a rapid and accurate evaluation of mass balances from large masses of data obtained in multiple surveys around a flotation circuit. The effects of changing process variables can be quickly evaluated with statistical confidence."

Jan 1, 1978

Analysing a flotation plant?s operating performance requires accurate metallurgical sampling. Typically, the aims of performing flotation plant surveys are to establish a consistent and accurate understanding of the operation of the circuit as well as to identify potential areas for improvement. Once the data have been collected, a mass balance is required to provide a consistent set of data from which plant performance can be assessed. Mass balancing is a key skill that many metallurgists and mining industry professionals are familiar with. Mass balancing reconciles the feed and product streams throughout the plant and is sometimes used for accounting of the material entering and leaving a system (eg. JKMetAccount). However more common reasons for mass balancing are: better understanding of the flowrates within a system, assessing the quality of data obtained from sampling and surveys, and providing a key dataset for assessing and improving the performance of unit operations. Mass balancing requires practice and experience to determine if the results are reasonable. There are various software programs available to assist in the mass balancing and understanding of processing plants. This paper is focused on JKMultiBal initially developed by Gay (1993), and then later commercialized by JKTech. JKMultiBal in its current release is for balancing data in both one and two dimensions, i.e. balancing data on either sized or assay bases as well as size-by-assay. With the recent trend in flotation circuits becoming more complex, JKMultiBal has been used to gain a better understanding of the processes occurring within the circuit. This paper provides a description of the program as well as several case studies where the two-dimensional mass balancing has resulted in improvements in plant performance.

Jan 1, 2009

Traditional loading systems have relied for many years on wheel and track loaders, power shovels, conveyor belts, and more recently, hydraulic front shovels. Each has certain advantages and disadvantages, depending largely on job conditions. Hydraulic backhoes were introduced several years ago as still another truck loading option. At first, buckets were small and backhoe use was limited to small earthmoving projects. Now, as backhoes are equipped with shorter sticks and bigger buckets to move larger quantities of earth. The "mass excavation" concept offers an excellent truck loading option in many situations. System Development and Advantages Mass excavation was first used successfully by contractors in site development and highway construction, loading both on-highway and off-highway trucks. Operators were able to take advantage of the excavator's ability to work in poor underfoot conditions and dig in tough materials. The technique was first applied to mining in the kaolin mines of central Georgia. Kaolin, one of the world's most abundant minerals, is used to make paper, paint, ceramics, rubber, plastics, and many other products. Since kaolin clay is slippery when wet, load-out with wheel loaders is very inefficient. Also, kaolin is found in very spotty deposits due to the weathering process involved in its formation. Consequently, draglines are also impractical for kaolin mining since they don't have the mobility to move between clay deposits. In response to these mining conditions and kaolin characteristics, mines began using excavators in mass excavation as early as 1976. They have developed a system that takes advantage of the excavator's •high penetration and breakout force in tight materials; •fast-cycling hydraulics; •ability to work in poor underfoot conditions; and •mobility in comparison to draglines. This loading method has been successful in Georgia's kaolin mines and now can be applied to other mining operations as well. The excavator's ability to carefully dig spotty ore deposits is an asset at uranium mining operations. Any mine can take ad-

Jan 3, 1982

By P. R. Scheller

Reactions between metal and slag are of decisive importance for metallurgical processing. In the continuous casting process, where metal solidifies at the metal-slag interface, reaction products can cause surface defects, with quality damages of the cast products as a consequence. These investigations are focused on the contact area between casting flux and metal. The samples were taken from the mould during the casting process. The results show that layers up to approximately 300 nm from the slag/metal interface represent a specific reaction space. Sodium in this layer on the slag side plays a specific role, which is explained on the basis of experimental results. The transport of oxygen near the interface and the charge transfer is explained on the basis of the ionic character of the slag. The mass transfer rate of oxygen crossing the interface is calculated from measured data, and conclusions concerning the stability of the interfacial tension are made. Keywords: mass exchange, metal slag reaction, interface, ion transport, oxygen transfer, oxidation, casting slag, flux, sodium

Jan 1, 2004

The Estonia Mine, operated by Eesti Energia Kaevandused AS (EEK), the mining division of the Estonian Government's energy company, Eesti Energia AS, is a large room and pillar mine located in Eastern Estonia. It supplied approximately 4.9 Mt/a of oil shale in financial year ending 2010, which is used for electricity and shale oil generation. Due to the Estonian climate and the power station's need for a dry product, Estonia Mine has developed a planning system where short-term demands of the power stations dictate the schedule. Such an arrangement requires maximising flexibility and consequently the mine only operated for 199 days in 2010.The footprint plays a key role in achieving mass mining rates as the 12 km + 12 km lease area allows the mine to split into mining districts that act independently of each other and maximise the number of work areas available. Coupled with this, the geology of the mine is relatively homogenous over the mining lease, simplifying the mine design and production methods.During the last five years, EEK has commissioned a comprehensive conveyor network. Consequently, material handling only requires loaders to dump oil shale direct from the production face onto a chain conveyor that feeds the conveyor network that ultimately supplies the enrichment plant. Dewatering is also of critical importance, with 90 Mm3 of water discharged annually. EEK uses large trenches up to 3.5 m deep and 1.2 m wide to channel groundwater. Track mounted mechanical saws are used to cut slots in the floor, to control overbreak in blasting these trenches.EEK has introduced Salzgitter Machinenbau GB 280S auger drills that drill three 280 mm diameter holes simultaneously. This has improved void hole drilling times and reduced the drilling required for each face, reducing the overall production cycle time. Blasting activities have also been decoupled from other mining activities negating the need for firing times through the use of small independent firings, required to control vibrations.

Dec 6, 2010

By J. K. Brimacombe, F. D. Richardson

Studies have been made of interfacial turbulence occurring in reactions in which indium, zinc and cadmium, dissolved in amalgams, are oxidised into aqueous phases by ferric or mercurous ions. It has been found that this turbulence occurs when the concentrations of oxidising ions in the aqueous phase become sufficiently great. These concentrations are the same as those at which the mass-transfer coefficients in the aqueous phase increase sharply when both phases are stirred by gas bubbles rising through them. The effects of gas flow rate on the mass-transfer coefficients for zinc in mercury have been established, and the oxidation of pure mercury has been investigated. The relevance of the work to processes involving liquid metals is discussed

Dec 1, 1971

By E. Barsky

"Theoretical notions of mass transfer processes in cascade separation of bulk materials have not been adequately developed as yet (Barsky M.D. 1980). In this paper we have derived linear difference equation of the second order for a cascade with initial feeding top any stage proceeding from mass exchange regularities. The solution of these equations has allowed us to determine quantitative dependences of the material distribution over the cascade stages, which serve the basis of the method of predictive estimate of the process results. We have also examined the model of the material distribution coefficient formation at individual cascade stages.We have checked the adequacy of this model to experimental data. It is noteworthy that the obtained theoretical results are phenomenological, i.e. they can be applied to the processes of separation of different physical nature, e.g. of isotopes, liquid media, gases, etc.INTRODUCTIONRecently, the processes of cascade separation of bulk materials have found broad application in various industries. They ensure a high separation efficiency and high-quality composition of final products. Although specialized literature gives serious attention to cascade processes [1], their theory is not sufficiently developed yet. To achieve optimal organization of this process, we examine mass transfer in a cascade gravitational classifier with stages of identical design."

Jan 1, 2016

By Ramani R. V

There are many important physical phenomena in mine ventilation which involve transient mass and heat transfer processes. These processes have been modeled using differential equations often solved with close form solution. Within the last two decades, many numerical schemes have been devised to solve these mass-transfer problems. Much of this effort has focused on variants of the finite element and finite difference methods. In this paper, the solution of mass transfer models in mine ventilation by finite difference technique is discussed. Specifically, the computational effort and the accuracy of the solution are presented.

Jan 1, 1988

By Sandin E. Phillipson

The Mine Safety and Health Administration?s Roof Control Division responded to a massive ground failure at an underground marble mine in northern Georgia. The ground failure occurred in a benched area that had been abandoned in the early 1990s. It involved the failure of an estimated 19 pillars, in conjunction with a massive roof fall of an undetermined extent that was at least 60 ft high. Pillars in the benched area had been mined nominally 40 x 40 feet to a height of 50 feet ft, with 40-ft mining widths, although it seems likely from measurements and accounts from mining personnel that pillars were even smaller than intended. Average overburden depth was 500 ft, and the rock mass was characterized by three sets of intersecting geological discontinuities that not only defined very blocky conditions that were conducive to pillar spalling and volume loss, but they were also adversely oriented in such a way as to cut diagonally through pillars at a high angle. The recently released S-Pillar software from NIOSH was used to determine pillar safety factors in the failure area and indicated that pillars were undersized for the conditions. S-Pillar was used to assess pillar stability in active areas to evaluate the likelihood of experiencing similar collapses in the future. S-Pillar correctly identified the pillars in the bench area as having been at risk for instability, in light of their ultimate failure. A two-dimensional finite element model was used to evaluate the possible stages of failure and to assess the likelihood of continued failure in the benched area that might affect more distal mine openings. Although stone mines are generally characterized as having more forgiving conditions than those found in softer sedimentary rocks, the failure shares some characteristics with domino-style massive pillar collapses experienced in coal and trona mines. First, the minimum dimension of the benched area approached 350 ft; second, the pillars defined by benching had small width-to-height ratios (0.8), and; third, the pillar failure area was slightly greater than 4 acres. The massive pillar collapse and associated roof fall illustrate the importance of incorporating pillars that have been properly sized for depth and for the final bench height. Furthermore, the designed dimensions must be adhered to during the actual mining process. Finally, the importance of understanding the geological conditions and their effect on the rock mass is critical.

Jan 1, 2012

One of the biggest challenges for the mining industry is, quite literally, how to the keep the equipment rolling. There is a limited amount of companies that manufacture the 4.2-m- (14-ft-) tall, 5,670-kg (12,500-lb-) tires needed for the haul trucks that are approaching the 362 t (400 st) pay-load class, such as Terex?s MT6300-AC and Caterpillar?s 797F. And, while these might be the most impressive tires, they are not the only ones that are in short supply. The shortage of tires has affected just about every kind of machine, and the demand for those tires has been extremely high for more than five years with no end to the shortage in site. American Tire Corp. (www.americantire.us) is one of the companies working to supply tires to the mining industry. It announced that the demand for its 1.6-m (63-in.) tires that have a zero failure history globally is great and that the company signed contracts worth more than $500 million at MinExpo. A relatively new player in the field of large tires for mining equipment is Titan Tire Corp., (www.titan-intl.com), a subsidiary of Titan International Inc. It shipped its first 1.6 m (63 in.) off-the-

Jan 1, 2008

By Chris Uys, Johan Labuschagne

Sishen Iron Ore Mine, located in the Northern Cape, is an open cast mine excavating approximately 100 million tons rock to produce 25 million tons of high grade haematite ore per annum. The size of the pit is approximately 12 km ‘1.5 km ‘150 metres. The final depth is planned to be around 350 metres. It is hard rock mining operation employing the conventional truck and shovel method. Mining benches are twelve and a half metres in height and are divided into blasting blocks, each consisting of approximately 300 000 tons of material.

Jan 1, 1997

By M. C. Pietrobon

"A common problem in assessing the applicability of laboratory flotation data is determining how closely conditions during laboratory testing of the ore matches those that occur in a plant. Typical laboratory based investigations do not consider factors such as water quality, pulp temperature and the pulp chemical environment during both grinding and conditioning on laboratory flotation performance. Disregarding these important factors may seriously compromise the validity of conclusions based solely on laboratory experimental results. Furthermore, considering these factors during the flowsheet design stage in the development of new ore resources may reduce financial risk associated with building a new plant. A methodology is described here which embraces these factors in an attempt to bridge the gap between laboratory and plant. The methodology should be considered a vital first step in all laboratory based experimental programs and has implications for fundamental studies on the flotation behaviour of single minerals. This methodology increases the confidence with which findings of laboratory based studies can be transferred to the plant and highlights the effect of variables which may be accessible to manipulation in the plant.Pulp chemical parameters, such as pulp dissolved oxygen demand, pH, Eh, mineral oxidation and solution composition were manipulated in the laboratory by adjusting the duration of gas purging after grinding, the process water input and pulp temperature. A key finding was that matching of the laboratory pulp chemistry to the plant pulp chemistry was dependent upon oxidation after grinding. In turn, mineral oxidation was dependent upon the pH in grinding, water composition and the duration of oxygen purging after grinding. The methodology closely matched the dissolved oxygen demand and surface chemistry of laboratory and plant pulps."

Jan 1, 2002

By Craig M. Sams

"Variable speed mill popularity has grown during the last 10 years with almost every SAG mill fitted with a variable speed drive. In addition, many of today’s larger ball mills are running a variable speed drive for various reasons. In the past, motor suppliers, especially gearless drive suppliers, have considered mills as being an overpriced rotor. This has lead to mills demanding more torque than what a motor can generate. This paper identifies the reasoning of what has happened and how to mitigate the mismatch in the future. Technical and economical issues are discussed, allowing mill operators to make an informed decision of which path to take in deciding what motor to put on a mill.INTRODUCTIONIt has been generally accepted in the past that motor suppliers and specialists in engineering companies understand the workings of a SAG mill or a ball mill and although there is a fair knowledge within the industry, in regards to mill torque requirements, the advent of variable speed drives in ball mills has shown the writer that a complete understanding of the mill torque requirements are not fully understood.This paper will discuss two aspects of mill torque requirements, (operating torque at various charge levels and start up torque), and how it relates to specifying the proper motor torque requirements.MILL OPERATING TORQUEMill operating torque at a given charge level and density in a mill is essentially constant throughout the entire speed range. For the purpose of this paper, we will take a 38’x 22.5’, 20’ EGL, SAG mill and provide the power draw at variable speeds but fixed loading conditions.The power draw and mill torque of this mill at 8% ball charge and 28% total loading (Highlighted) is shown below as Figure 1, at variable speeds."

Jan 1, 2007

By J. P. Harrison, J. A. Hudson

ABSTRACT: Given the current widespread use of numerical models for rock mechanics and rock engineering, and the ‘Technology Road Map’ theme of the Congress, it is timely to consider the key issues in matching numerical models to rock engineering problems. The three main aspects are: a) identifying the mechanisms, variables and parameters that are relevant to a particular rock engineering project; b) establishing the content and mode of operation of a particular code; and c) deciding whether a particular code provides an adequate model for a particular project. We summarise the key issues, highlighting the concept of rock mass domains, validation and the ability to present the results of a modelling exercise. We include three compact numerical modelling case examples to illustrate the points. Finally, we also present a ‘roadmap’ to assist in the process of matching a numerical model to the rock reality. RESUME: Considérant l’utilisation généralisée des modèles numériques en mécanique des roches et en ingénierie au rocher ainsi que le thème du congrès ‘technology road map’, il est opportun d’aborder les problèmes essentiels liés au choix des modèles numériques dans les problèmes ingénierie au rocher. Les trois aspects principaux sont : a) l’identification des mécanismes, variables et paramètres appropriés au projet considéré, b) la définition du contenu et du mode de fonctionnement d’un code numérique donné, et c) la possibilité de décider si un code donné fourni un modèle approprié au projet considéré. Les points clés sont résumés en mettant en évidence les notions de “massif rocheux”, de validation et de capacité à présenter les résultats d’une approche numérique. Nous présentons trois brèves études de cas qui illustrent ces points. Enfin, nous proposons aussi un logigramme pour aider au choix d’un modèle numérique approprié pour représenter la réalité d’un problème au rocher. ZUSAMMENFASSUNG: Im Licht der augenblicklich weitverbreiteten Nutzung von numerischen Modellen in der Felsmechanik and im Felsingenieurbau und unter Beruecksichtigung der Themenstellung des Congresses “Technology Road Map”, erscheint es an der Zeit die Kernprobleme, die sich mit Fragen der Abstimmung der numerischen Modele mit Hinblick auf ihre Anwendbarkeit im Felsingenieurbau beschaeftigen, naeher zu befassen. Die drei wesentlichen Aspekte sind: a) Identifizierung der Mechanismen, Variablen und Parameter, die fuer ein bestimmtes Projekt im Felsingenieurbau von Bedeutung sind; b) die Festlegung des Inhalts und des Modus Operandi eines bestimmten Codes, und c) die Entscheidung darueber ob ein bestimmter Code ein hinreichendes Model fuer ein bestimmtes Projekt ergibt. Wir fassen die wichtigsten Punkte zusammen, heben dabei das Concept der Gesteinsmassen Domaenen (Rock mass domains) hervor, und erarbeiten ihre Beurteilung und die Moeglichkeit diese zu praesentieren in den Ergebnissen einer Model Uebung. Wir schliessen drei Beispiele von numerischen Modellen ein in denen diese Punkte erlauetert werden. Zuletzt schlagen wir eine “Roadmap” vor, um dem Prozess der Abstimmung des numerischen Modells mit der “Felsmechanischen Wirklichkeit” zu unterstuetzen.

Jan 1, 2003

By Claude Bazin

"The On-Stream Analyzer has proven and is still proving to be an extremely valuable tool for the control of flotation circuits. This paper examines the possibility of going further in the utilization of the provided assays using a real-time material balance program to calculate the distribution of solids within the streams of a circuit. Mass balance results can be used for process control, process analysis and detection of sensors malfunction. Examples from 3 Canadian and one American concentrators are presented to illustrate the potential of the method.IntroductionThe On-Stream Analyzer (OSA) has gained, over the last two decades, the status of key instrument for the control of flotation plants. The OSA is currently used in almost every sulphide flotation plant and its application in oxide mineral flotation plants is increasing. On-Stream Analyzers provide X-rays pulse counts that are converted into assays using multilinear regression programs. The estimated assays are passed to the flotation operators, via an operator console, or to control loops of various degrees of complexity (McKee, 1991) implemented in a Distributed Control System (DCS). Operators and/or controllers take control actions based on these assays to ensure an adequate operation of the plant. Although this application pays off very rapidly for the purchase of the OSA, one should ask: Could we get more from the OSA?Some plants reported the use of OSA measurements to process optimization (McKee, 1991). The optimization strategy is based upon economic indices calculated from metal recovery using the 2 product formulae (Wills, 1992, page 124). The 2 product formulae is however sensitive to measurement errors in the assays and can lead to considerable scatter in the calculated indices (Hodouin et al., 1993). Least squares algorithms provide an interesting alternative to filter out measurement errors (Spring, 1992). This paper presents a recursive least squares algorithm that estimates in a real time fashion the mass flow rates in a flotation plant using OSA and mass flow measurements. The algorithm is robust and its recursive structure is not demanding in terms of computation time and memory. The method was tested with data from four concentrators. Results of the material balance algorithm prove to be useful to detect sensors malfunction and may potentially be used for the control and optimization of flotation plants.The paper is divided into two sections. The first section gives a description of the proposed method. The mathematics of the method are kept to a minimum, since the detail is presented elsewhere (Bazin and Franklin, 1994). The second section shows the application of the method to different flotation plants."

Jan 1, 1995

By Schonert K

The throughput of a HP-roller mill is determined by the acceleration of the feed material within acceleration zone and introduction into the compression zone. Both processes depend on the internal friction and the friction of the material on the roller surface which is influenced by the surface structure and adhering of material. The particle motion in the acceleration zone has been investigated with high speed videos showing the effect of the centrifugal force and the slip on the roller surface. The throughput is discussed in the term of the specific throughput which can be approximated by a linear function characterised by the throughput number and the throughput elasticity. The throughput number is a helpful measure for investigating the influence of the feed material and the surface structure. The experimental results give some guidance for chosing an optimum corrugation profile.

Jan 1, 1993

By Thomas M. Barczak

Secondary roof support systems play a vital role in preserving the safety of underground mine workers by preventing the unintentional collapse of the mine roof. Hundreds of thousands of standing roof supports are constructed each year in underground coal mines. Historically, wood and concrete cribs and timber posts have been used for secondary roof support. These support constructions require the handling of heavy and bulky materials, causing numerous injuries to the mine workers and resulting in more than 40,000 lost workdays in the past 9 years. Since 1993, various alternative support technologies have been developed. These new support technologies not only provide superior roof support, but many also have significant material handling advantages by using smaller and lighter weight materials, fewer components, mechanically installed support systems, and pumpable support systems. The National Institute for Occupational Safety and Health (NIOSH) has conducted extensive material handling research and has developed recommended lifting thresholds to reduce material handling injuries. An analysis of roof support construction reveals that conventional support materials used in wood and concrete cribbing exceed the recommended lifting thresholds, while the engineered support systems are closer to the recommended weight thresholds. Finally, recommendations are made relative to proper lifting techniques and material handling practices to prevent injury to mine workers when constructing roof supports in underground coal mines.

Jan 10, 2000

By J. F. Oyler

One of the most important applications of material handling engineering technology occurs in iron ore pelletizing plants. Although the total installed value of material handling systems in such plants, as well as in many other modern process plants, is a relatively small percentage of the value of the plants (5 to 10% perhaps), their significance is considerable. Indeed this significance equals that of the more expensive process equipment in the plant, since the failure of any component of any system causes, shutdown of the entire plant. It is because of this increased consequence of failure that design criteria for process plants are more conservative than for conventional material handling applications. A second unique characteristic of material handling technology applied to process plants in general is the variety of materials being handled and the wide range of problems this variety presents. One extreme encountered is the stickiness of filter cake. Here the entire design effort is aimed at keeping the material flowing through transfer points and preventing it from building up on components. Another extreme is the conveying of fired pellets - hard spheres with a very flat angle of repose and abrasive characteristics severe enough to make conventional wear plate materials impractical.

Jan 1, 1976

By John W. Fredland

Many accidents occur in the mining industry as a result of the instability of material during handling and processing operations. Accidents due to dump point instability at stockpiles, and at spoil or waste piles, for example, occur with alarming frequency. Miners must be trained to be better aware of these hazards. Information on safe working procedures at stockpiles and surge piles is provided. Mine operators must review their training and operating procedures regularly to ensure that hazardous conditions are avoided.

Jan 1, 1993

By Scott R. Giese, Lucas Dix

"Recently, there is a growing trend in the aluminum foundry industry to develop· improved melting and casting processes to produce lightweight, high-strength, and thin-wall castings for the aerospace and transportation industry. One casting method that has demonstrated industrial potential for producing clean, thinwall castings is vacuum-assisted, counter gravity casting. This process has the benefit of expeditiously and quiescently filling the casting cavity with a dramatic reduction of furnace related and reoxidation inclusions. Additionally, lower casting temperatures can be utilized because the application of vacuum during filling overcomes the resistive internal cavity pressure resulting in a finer, as-cast microstructure. The paper explores and discusses the processing and design conditions in improving the mechanical properties of A206 casting alloy using the vacuum assisted, counter gravity casting process.IntroductionThe process of metal casting has been, over the centuries, the chief production technique for components used in everyday life. However, the casting process has certain inherent problems that can adversely effect the properties of a finished product. These problems are directly attributable to the gravity pouring technique used to transfer molten metal to the mold and to fill the mold cavity.A recent study for the American Metalcasting Consortium demonstrated reoxidation products created during the casting process are five times more likely to cause product rejects than entrapped molding aggregates, melt slag, or furnace lining materials1. Formation of nonmetallic and intennetallic compounds are more prevalent in aluminum and titanium alloys because of their reactiveness to furnace lining materials and their high affinity for oxygen. The incorporation of undesirable compounds can be attributed to four areas of the casting operation. The sources are charge materials, melting technique, pouring methodology, and casting cavity filling. Aluminum foundry men have addressed the first three areas by selecting high quality melt stock, non-stirring melting equipment, and the use of porous ceramic filters. Unfortunately, little consideration is given to the last, but yet tl.1e most important part of the casting operation, the actual fill mg of the casting cavity.The generation of inclusion defects during the casting process can be described by two methods. As the liquid metal flows in the gating network and casting cavity, abrupt changes in flow d1rect10n cause the metal to become agitated. This mechanism liquid metal splashing, is generated by instability of the advancing liquid metal interface from turbulence effects as schematically illustrated in figure 1a2 . The surface oxide film bursts under high velocity conditions, folds over, and becomes dispersed and incorporated in the bulk liquid. Additionally, fragmented liquid droplets ejected from the disturbed interface oxidize and become dispersed in the metal stream."

Jan 1, 2000