Search Documents

By J. J. Kim

In a study of buildup in a green-feed reverberatory furnace, it was determined that approximately half of the crucible area has a moderate buildup of 25 in. or less, and that the area extending 45 f t from the burner end of the furnace has only a slight buildup. The false bottom, averaging 6 in. in thickness, was found to cover the entire crucible area. The stratification of Cr2O3, A12O3, and SiO2 occurs throughout the slag layer; this phenomenon is the principal reason for the formation of the false bottom in the green-feed reverberatory furnace. The false bottom is caused primarily by the settling of FeO*(A1,Cr),03 aggregates, which are derived mainly from returned converter slag. The false bottom can be transformed to a solid permanent buildup in the furnace, particularly when the matte is completely withdrawn. From the material balance of Al2O3 and Cr2O3 around the furnace, It is estimated that the average rate of aluminum chromite accumulation is approximately 0.3 ton per day or less.

Jan 1, 1976

By Stefan Mutl

By the term sorting one understands the separation of a material mixture into two or several products according to different material characteristics in the preparation technology. The oldest sorting procedure, the handpicking, is less and less used to separate different material mixtures because of the considerable personnel expenditure and the strict hygienic regulations at the workplaces in Europe. In general, the manual assortment is able to solve separation tasks, which could not be solved by machines so far. The picking is one of the most selective processing methods. However, the progress in the computer and sensor engineering today makes the creation of automatic picking devices possible, whose efficiency goes far beyond the possibilities of handpicking. Therefore in the future automatic picking will not only be applied to already common applications like the waste glass assortment by colour, but also to the assortment of complex composed material mixtures e.g. old shreddered cars. Thus, the automatic picking, supported by high performance computer systems and new sensor technology, gains more and more significance within many applications of waste processing. However, in order to apply automatic sorting procedures for different material streams within recycling processes, a substantial research and development need still is existing. The desire of the users is to sort more extensively in order to achieve products with bigger material varieties, In addition, the requirement of the accuracy of recognizing and sorting out of single grains rises, This lecture illustrates the procedure of the hand picking before it was employed in waste processing. Thus, a renaissance of a long forgotten method is happening in terms of today's state of the art procedure "automatic picking". Possible technologies, fields of application and future potentials of this technology are pointed out.

Jan 1, 2003

By Brian Balmforth

Solvent extraction has been applied successfully to the recovery of a range of metals from a variety of process streams. Particular benefits are derived from the use of new highly selective extractants. These will be exemplified by case studies including the recovery of zinc from sulfate and chloride media using ZENECA's new extractants which show a very high selectivity over iron and the recovery of antimony and bismuth from copper refinery electrolytes using ZENECA's strong and selective extractant DS5834. The conclusions are drawn that solvent extraction is a cost effective technology for metal recovery and recycling, even when used to treat relatively small volume/low concentration process streams, and that it can provide the essential enabling components of a larger system for waste minimisation programmes.

Jan 1, 1996

By Robert G. Telewiak

During the past few years, the international nickel industry has taken some important steps to respond to conditions of global production overcapacity, relatively flat demand and low prices. These include the implementation of major cost reduction programs by most producers, closure of some high cost operations and the establishment of the Nickel Development Institute. Consumption in the western world increased by about 7 per cent in 1987 compared to 1986, supplies were less than demand, inventories were low, and late in the year prices increased dramatically. By year-end 1987, most producers were operating at, or close to, their effective capacities. Nickel demand is expected to increase to the year 2000 at an annual rate of about 1.7 per cent, led by increasing usage in the stainless steel sector. If this rate of growth is realized, then there will be need for the effective capacity of the nickel industry to be increased. Producers in Canada have put considerable effort in the past few years into reducing costs - with impressive results. Mining costs, in particular, have been reduced dramatically. While production is still less than the peak years of the yearly 1970's, it has increased substantially from levels of the early 1980's. Over the next few years, production is expected to increase slowly. A key determinant of the level of production in the 1990's could be the effect that regulations on sulphur dioxide emissions have on limiting production.

Jan 1, 1988

By Alan James Parker

Chalcopyrite can be converted to pure copper by the following five steps. An exothermic sulfation roast of chalcopyrite; leaching of cupric sulfate from the calcine with dilute acid; precipitation of Chevreul's salt and or other copper sulfites with ammonium sulfite or precipitation of crude copper by autoclaving ammonium sulfite and cupric sulfate at 150°C; dissolution of Chevreul's salt, or the crude copper, as cuprous sulfate, using acidic cupric sulfate in an acetonitrile-water solution as oxidant; precipitation of pure copper by thermal disproportionation of the cuprous sulfate solution. Acetonitrile and acidic cupric sulfate solution recycle. Advantages over conventional processes include lower cost, lower energy consumption as waste heat, sulfur removal as ammonium sulfate or gypsum, rather than as sulfur dioxide and rapid throughput. The net reaction is: 2 CuFeS2 + 4H20 + 15/2 02 + 8 NH3?2Cu + 4 (NR4)2 SO4 + Fe2O3

Jan 1, 1976

By S. P. Vanka, B. G. Thomas, S. Sivaramakrishnan

"During the continuous casting of steel, transient 'flow events can be very important to the generation of quality problems, such as shear entrainment, surface level fluctuations and bubble entrainment. These phenomena can be better predicted using Large Eddy Simulation (LES) models than the K-E models, which have been extensively studied. In this work the turbulent transient flow in a water model of a continuous caster has been both simulated using LES models and measured using Particle Image Velocimetry (PIV). The results of PIV and LES have been compared to validate the LES model and also to enhance the understanding of the prominent transient features of the flow field. The flow near the top surface in PIV was found to contain periods of 5-1 Os when the velocities were three to four times their mean values, which is of importance to the shear entrainment of the liquid flux. In both simulation and PIV the upper roll structure is found to change chaotically from a single large recirculation structure to a set of distinct vortices. The inlet swirl in PIV is seen to persist more than halfway across the mold, causing a characteristic staircase velocity vector pattern when viewed in a plane parallel to the wide faces. The lower rolls in PIV are significantly asymmetric for very long periods of time (- I-hour) and go through a repeating sequence of features. One of these features involving a short circuit between the upward and downward flow in the lower roll is also seen in the simulation and is of significance to bubble entrapment.IntroductionThe turbulent flow through the nozzle and in the mold of the continuous caster has been studied extensively using computational Reynolds Averaged Navier Stokes (RANS) models of turbulent flow [1,2]. Defects, such as internal inclusion of liquid fl4x through surface shear, pencil pipe defects due to bubble entrapment, surface defects due to level fluctuations are caused by transient flow phenomena oc'curring in the mold. These phenomena can be better predicted using transient Large Eddy Simulation [3] or ""LES"" models."

Jan 1, 2000

By Egil Krystad, Merete Tangstad, Kai Tang, Gabriella Trenell

"A new UMG-Si refining concept has been proposed in this paper. Liquid Si and another liquid metal phase (Me) were separated by the molten slag spontaneously based on their different physicochemical properties. Two metal/slag distribution reactions are able to occur in the system simultaneously. Experimental investigation on the B and P distribution equilibria among Si/slag/Me was carried out in the graphite resistance furnace under the inert atmosphere. The B and P doped silicon was in equilibrium with either the SiO2-CaO or SiO2-CaO-l0%CaF2 slags at 1600 and 1650°C, respectively. The pure Fe and Cu metals were used in order to ""capture"" B and P in the slag phases. B and P in samples were determined by the ICP-MS. B and P distribution coefficients for the Si/slag/Me equilibria were compared with the theoretical calculations. The possible kinetic barriers for the Si/slag/Me three-layer equilibria have been analyzed. The metal contaminations in the three-layer equilibria were also examined. The present study provides a new approach for the refining of MG-Si.IntroductionMost of the impurities in the metallurgical silicon (MG-Si) are rather ""dilute'', ranging from few to several thousand ppms. However, the target impurity levels for upgrade metallurgical silicon (UMG-Si) feedstock are often far below the level of ppm. In such extreme case, the conventional refining technologies face the great technical challenges. B removal by slagging has already been implemented industrially. Removal of Pis frequently carried out by the vacuum distillation. The separating treatments for B and P removals lead to increase refining cost, total CO2 emission and longer pay-back time of solar cell module.It is well known that the phosphate capacities of slags in equilibrium with liquid silicon are extremely lower. However, if we introduce another phase that possesses high ""phosphor capacity"" into the Si/slag system, one would expect to continuously remove P from Si through molten slag to the new ""phosphor capturer"". The so-called three-layer refming concept is thus proposed in the present study. Liquid Si and another liquid phase (Me) are separated by the molten slag spontaneously based on their different physicochemical properties. Two metal/slag distribution reactions are able to occur in the system simultaneously."

Jan 1, 2013

By Cesar Garban

The samples of the washings are characterized of the ore for physical, chemical and metalurgical properties for representivity. the optimum values are determined for the recuperation of gold of the washings of the gold ore from the zone El Callao, Venezuela by acidic solutions of thiourea of the following parameters: Ph , contact time, composition of the liquid mixture, ratio of solid to liquid, grade of liberation, in an apparatus designed for contacting solid powder of the mineral with liquid solution of thiourea mantained in suspension by bubles of compressed air. The experimental results showed that the velocity of lixivation in primier time of (20) minutes is rapid and practically 2/3 part of the reaction is completed. For experimental design of parameters, the optimum conditions of recuperation of gold determind in this experimental investigation are:

Jan 1, 1999

By Levente Becze

The zinc bottom dross is obtained as a by-product from hot dip galvanizing operations and contains at least 92 % zinc, which is recycled mainly by pyrometallurgical processes. The overall costs of such a treatment, however, are often relatively high, and the very strict environmental protection regulations might also cause additional difficulties for the high temperature procedures in the future. Therefore, a novel hydro-electrometallurgical process has been developed in our laboratory, by which one can reclaim zinc of high purity, while attempting to achieve low energy consumption and making allowances for all the environmental restrictions. This aqueous processing technique, based on electrolytic refining in a cell equipped with anion-exchange membrane to separate the anode and cathode compartments, allows the indirect purification of the anolyte. Zinc is dissolved from the anode along with some of the impurity elements, while the intermetallic compounds (Fe-Zn, Fe-AI- Zn) are accumulated in the anode slime.

Jan 1, 2003

By Desmond Cook, Cynthia P. Dogan, Dale Govier, Gary Korth, James C. Rawers

"The microstructural and mechanical properties of iron powders processed in an attrition mill are highly dependent upon the processing environment. Extended high-energy ball milling in argon resulted in nanocrystalline bcc-Fe microstructure, while processing in nitrogen or alloying with carbon resulted in both nanocrystalline bcc-Fe and bct-Fe microstructures. In this U.S. Bureau of Mines study, both X-ray diffraction and Mössbauer analysis were used to show that the infused carbon and nitrogen concentrations were predominantly on the grain boundaries. Explosive consolidation resulted in a fully dense product with microstructure similar to the microstructure of the mechanically processed powder. Hardness values of the compacts correlated highly with grain size, independent of the processing conditions.IntroductionNanocrystalline materials have been theorized and shown to have unique properties that make them interesting for material scientists and promising for material engineering applications. Several conferences have been devoted solely to the techniques of preparing nanostructured materials have been held(l,2,3.). To date, most techniques used to generate nanocrystalline materials produce only a few grams of particles(1). High-energy ball-milling, however, is an efficient and effective technique that can produce substantial quantities of nanocrystalline materials with homogenous chemical composition that cannot be produced by conventional processing techniques(3).Applications and material evaluations of nanocrystalline materials require several grams of solid samples. Because nanocrystalline microstructures' are thermodynamically unstable, producing consolidated macro structures of nanocrystalline material has proved to be very difficult(2,3). In particular, retention of the nanocrystalline microstructure during consolidation of nanocrystalline particles has proven to be a problem. In contrast to many consolidation techniques that require elevated temperatures, explosive compaction uses a shock wave for consolidation( 4,5)."

Jan 1, 1994

By H. Y. Sohn, M. Pérez-Tello, P. J. Smith, K. M. Riihilahti

A three-dimensional computer model for the Kennecott-Outokumpu flash converting process for copper matte is presented. The model incorporates the transport of momentum, heat and mass, and reaction kinetics between the gas and particles in a particle-laden turbulent gas jet. The k-E model is used to describe gas-phase turbulence in an Eulerian framework. The particlephase is treated from a Lagrangian viewpoint which is coupled to the gas phase through the source terms in the Eulerian gas-phase governing equations. Matte particles were represented as Cu2S·yFeSx. Based on experimental observation, the oxidation products were assumed to be Cu2O, CuO, Fe3O4, and SO2. A reaction mechanism involving the external mass transfer of oxygen to the particle surface and diffusion of the oxygen through the oxide layers of CuzO/Fe304 and CuO/Fe3O4 is proposed. Predictions of the mathematical model were compared with the experimental data collected in a large laboratory furnace. A reasonable agreement between the model predictions and the measurements was obtained in terms of fractional conversion of the oxidation reactions and of sulfur remaining in the reacted particles. The simulation of an industrial flash converting furnace was also performed. Higher oxidation rates, more even distribution of particles, and thus a more efficient use of the reactor volume are predicted with a burner having a distributor cone, than with a single-entry burner.

Jan 1, 1999

By S. N. Ojha, Gouthama, R. K. Mandal

"Spray deposition processing of liquid immiscible alloys, based on AI-Pb and Cu-Pb systems have been investigated. Process variables have been optimized to generate uniform dispersion of ultrafine second phase particles in· the matrix of the primary phase. Mechanisms governing the microstructural evolution in this synthesis methodology are discussed. Results of the present investigation are used to demonstrate that rapid solidification of atomized droplets initially provides a high nucleation rate of the second phase during the liquid phase separation. The solid matrix phase nucleates at a later stage and grows with a high velocity enabling trapping of the dispersed phase. Subsequent alteration in the size and size distribution of the second phase is suggested to arise further during deposition of high velocity droplets. Possible reasons for this effect are brought out. Microstructural features of atomized powders and spray-deposits of these alloys are presented to support the present approach of synthesizing composites. IntroductionIn recent years, synthesis of advanced materials with metastable microstructures and ultra fine dispersion of second phase particles has led to the development of novel techniques of materials processing. The other requirement of such a processing methodology is to achieve near-net shape geometries in less number of processing steps. The solidification processing method via spray deposition route qualifies these criteria. This process primarily involves atomization of a melt by high energy gas jets into spray of droplets which are subsequently propelled by the kinetic energy of the gas towards a deposition substrate to build-up a thick high density preform. Although the concept of spray deposition was pioneered by Singer [1] in the early seventies, the fundamentals of the deposition process has been understood only recently [2-4]. Several benefits of spray deposition processing are by now well established [5,6]. Rapid solidification effects are inherent in spray deposition processing due to high heat exchange rate at the droplet-gas interface and also on the deposition surface by force convection. Since considerable heat extraction occurs from the atomized spray prior to droplet consolidation, a relatively low processing temperature is ensured in this process. As a result deleterious coarsening phenomenon can be minimized. In addition, controlled maneuvering of the substrate provides the capability of the process in producing disc, sheet or tube shape preforms of difficult to work materials in a single operation. A significant improvement in properties, resulting from microstructural homogeneity and refinement in grain size induced by spray deposition has been reported for several alloy systems [7,8]."

Jan 1, 1994

By Sebastião Ribeiro, Leonardo Carneiro Sardinha, Alline Sardinha Cordeiro Morais, Carlos Maurício Fontes Vieira, Sergio Neves Monteiro

The glass tube of spent fluorescent lamps is contaminated with mercury, which might be a serious hazard in the case of conventional recycling by melting with other glasses. A possible solution could be its incorporation into a clay body to fabricate common fired ceramics such as bricks and tiles. The objective of this work is to characterize a type of fluorescent lamp glass waste to be incorporated into a clayey ceramic. The characterization was performed in terms of wettability tests to evaluate the interaction between the surface of the clayey ceramic and glass waste as a function of the firing temperature. The results showed that the contact angle decreased with increasing temperature, reaching a value of 79°, at a temperature of 1100°C, but not sufficient to completely wet the ceramic. However, compatible chemical composition and reduction of porosity by the flow of soft glass waste between the clay particles favor the consolidation of the ceramics structure above 900°C.

Jan 1, 2015

By Kazuhiro Nisizono, Takashi Kuroki, Nobuyuki Mori, Keisaku Ogi

"The improvement of critical current density and mechanical property is essential to the practical application of the superconductive oxides. A solidification technique is developed in this study for the growth of columnar or single Y123 crystal with finely dispersed Y211 and Ag particles through a monotecto-peritectic reaction of Li.(Ba-Cu-Ag-0 rich liquid)+ Y211-+ Y123+Lz(Ag rich liquid). The large solidification rate and temperature gradient can refine the Ag and Y211 particles. But Ag-free region is widely formed along the crystal growth direction in the central part of the Y123 cell. Since this phenomenon occurs because of the cellular growth, it is necessary to narrow Y123 cell width or to grow Y123 crystal with flat freezing interface. By controlling crystal growth direction using a seed crystal, Y123 columnar or single crystal with uniformly dispersed fine Ag particles are obtained. The· behavior of Ag and Y211 particles during the solidification of Y123 crystal is simulated by the peritectic growth model of faceted Y123 crystals.IntroductionY-Ba-Cu-0 superconducting oxide (YBCO) with high critical temperature also has excellent superconducting characteristic under the high magnetic field. The critical current density (Jc) has been enhanced by ·using unidirectional-solidification technique, which can overcome the weak-link problem at grain boundaries and control the crystal orientation and the distribution of Y211 particles which give effective magnetic flux pinning centers in the Yl23 matrix [1-4].In addition to improving superconducting properties, it is necessary to improve mechanical property for the practical application. The silver addition to the melt-textured YBCO superconducting oxide has been reported as the effective method to improve mechanical properties without degrading superconducting properties. Since the material properties are considered to be largely affected by silver particle size, form, dispersion state etc. in the Y123 phase, it is important to control Ag distribution in YBCO/Ag composites. Besides, it is also reported that the silver addition improve the corrosion resistance to the water and CO2 [5,6]. Nishizono et al. [7-9] have reported that the silver distribution in the YBCO/Ag composites is largely affected by the solidification condition, but yet some unclear points are also remained. YBCO has been also fabricated by the zone melting method [10-12], and Furuya et al. [5] and Hirano et al. [13] have reported the fabrication of silver added YBCO composites, but the distribution and behavior of Ag are not investigated in detail"

Jan 1, 2000

By J. R. Bhamidipati, Nagy EI-Kaddah

"Over the years, mathematical modeling has been established as a viable tool for design and analysis of melting and solidification processes. Modeling of electromagnetic casting and melting systems involves three coupled problems -- determination of the containment field, the shape of the meniscus, and the temperature field -- which have to be solved simultaneously. One of the difficulties in numerical simulation of these systems is in defining and gridding the solution domain as the meniscus shape of the liquid region is not known a priori. This paper describes a methodology for solving these coupled problems in a two-dimensional electromagnetic containment field. This method is based on the solution of the electromagnetic, free surface dynamic, and heat transfer equations in a fixed domain in computational space by mapping the physical space into an invariant computational space during the search for the equilibrium shape using a variable non-orthogonal coordinate transformation. This approach, also, eliminates the need for regridding the liquid and solid regions during phase change. This methodology is demonstrated for two electromagnetic confinement systems -- the Magnetic Suspension Melting process and the electromagnetic casting of aluminum. The model predictions are in good agreement with available data for these systems. It is suggested that the methodology presented here is likely to provide a useful tool in the design, analysis, and optimization of electromagnetic confinement systems. IntroductionDuring the past decade, mathematical modeling has emerged as a viable tool for the design, analysis, and optimization of casting and melting processes /1-2/. It is generally recognized that the usefulness of modeling to study the interplay between the various process parameters without resorting to extensive experimentation relies heavily on accurate representation of the physical phenomena in these systems. In modeling electromagnetic confinement casting and melting systems, it is important that the model address all aspects of electromagnetic field interaction with the metal, if accurate predictions are to be obtained /3/. The electromagnetic field affects the heat transfer phenomena primarily through joule heating. In addition, the electromagnetic forces play a key role in heat redistribution in the molten pool via melt stirring, and, in confinement systems, support the molten metal pool."

Jan 1, 1994

By R. Hurman Eric

The flow of molten slag and matte and the estimation of the mass of settled matte in electric furnaces is discussed and the action of the electric furnace as a settler is considered. Particle size of the concentrate and its grade are major contributors to settler dimensions. Discussion on the production capacity of the electric furnace with reference to operating current, power input, furnace dimensions such as width and length and electrode diameter are given. These discussions are based on considerations derived mainly from experimentally determined properties of the slag such as its resistivity, viscosity and liquidus temperature. Extension of the design and operating considerations to circular furnace symmetry is also discussed.

Jan 1, 1995

By Merete Tangstad, Gabriella Tranell, Jeff Kline

"Raman spectroscopy as an instrumental technique for the determination of silicate structure is widely accepted. This method was utilized for analysis of structural modifications associated with the addition of network modifying species. Silicate melts are described by the extent of oxygen bridging. Therefore, understanding the effect different oxides have on modifying the silicate structure will provide valuable information regarding the removal of boron from silicon in slag refining processes. Samples ranging from CaO/SiO2 = 0.56-1.2 were evaluated with and without varying concentrations of B2O3. The addition of CaO to SiO2 revealed Q2 species constitute 45% of the Qn species and addition of B2O3 to the high concentration CaO system results in an increase in Q2 and a decrease in Q0 . The disappearance of the Raman band associated to the three-member ring in the low concentrated CaO system indicates boron addition and destruction of the ring.IntroductionExtensive studies on the structure of silicate systems have been performed using Raman spectroscopy with consensus on the Raman shift associated with different bridging oxygen units [1]. Most of these studies are motivated by the lack of a fundamental understanding of the processes involved in petrology. Due to this, an opportunity exists to build on the knowledge obtained from these studies and conduct more research on slag structures relevant to industrial applications, as these systems are often compositionally similar to magmatic melts. Raman spectroscopic studies of industrial slags will provide insight into the possible structural modifications necessary to enhance removal of impurity species from the metal phase to the slag phase."

Jan 1, 2012

By John A. Thorpe, Gary R. Maxwell, Kirk M. Schall, Kenneth A. Brunk

"Gold ores containing oxide and some sulfide copper minerals incur significant cost penalties when leached with cyanide, because 4-5 moles of cyanide are consumed for each mole of leached copper. This equates to ~4 kg of NaCN per kg of leached copper. Total cost of excess NaCN consumption plus CN destruction can exceed $4.00 per tonne of ore. AuGMENT Technologies has patents pending on a process that can 1) economically recover 70-80% of the cyanide that is fed to a mill, 2) produce salable copper cathodes, and 3) reduce the need for a cyanide destruction facility. Standard equipment for resin processing and electrowinning are used to recover the cyanide and produce the copper cathodes. Steps have been taken to determine: 1) how to develop the resin kinetics, 2) how much savings are incurred at a mill site, 3) how the novel 'electro-elution' circuit achieves efficient operation.IntroductionWhen gold ores containing nuisance copper are leached with cyanide, a significant cost penalty is incurred because 4-5 moles of cyanide are consumed for each mole of copper that leaches. This is equivalent to the use of -4 kg of NaCN per kg of copper that leaches. The total cost of excess NaCN consumption plus cyanide destruction can exceed $4.00 per tonne of ore.AuGMENT Technologies has a patent pending on a process that can 1) economically recover 70-80% of the cyanide that is fed to a mill, 2) produce salable copper cathodes, and 3) reduce the need for a cyanide destruction facility. This ACCS (Au-.C.u-.C.N Solution) Process Technology involves the novel application of copper cyanide chemistry. Standard resin processing and electrowinning equipment are used to recover the cyanide and produce the copper cathodes."

Jan 1, 1997

By C. J. Beyke

The U.S. Bureau of Mines has developed a process for the selective recovery of Pb metal from a mixed Pb-Zn concentrate. The concentrate was leached in recycled H2SiF6 at atmospheric pressure and 95° C with a redox potential of 570 mV (Eh). Eighty-five pct of the Pb was selectively leached from the concentrate while 87 pct of the Zn remained in the residue. Pressure leaching of the concentrate was slightly less successful, with 78 pct of the Pb leaching from the concentrate and 80 pct of the Zn remaining in the residue. Pure Pb metal was successfully electrowon from a purified PbSiF6-H2SiF6 solution with 96 pct current efficiency and 0.69-kW.h/kg energy consumption.

Jan 1, 1991

By Li Xi

This paper describes how a back-propagation neural network is built for predicting the properties of engineering materials from their other properties, compositions or processes. A predicted result of the impact toughness of a 12Cr-2Ni-1.5Mo stainless steel from its tempering temperature is presented. Furthermore, the combination of this neural network for prediction is able to supplement the unpredictability of the expert system outside its domain of expertise. An integrated system may perform more intelligently and solve more complex problems as well as human expert, in some case even better.

Jan 1, 1992