Search Documents

By Paul D. Tomlingson

Part one of this two-part series covers conceptualizing and developing a maintenance information system. It explains what information is necessary to manage such a system. Next month, part two will discuss using this type of system with administrative information, including error and absentee reports as well as open work files. Introduction There is no doubt that quality, timely, and accurate information can make a solid contribution to maintenance cost reduction objectives. Yet, despite obvious benefits, methods being used to acquire information systems are proving costly, frustrating, and often unsuccessful. Those needing the information are disappointed. Mine managers, hoping to reduce costs with quality information, are instead seeing no system, no savings, and massive spending on information endeavors of questionable benefit. Options Any software vendor will happily sell you his program. Numerous mining organizations, having "gone to the cleaners" themselves, will gladly make their programs available to help cover huge trial and error development costs. Any package can only be a generalization. Any program bought from some other organization must undergo numerous modifications to make it work properly. In the process the initial cost is doubled. Many mining maintenance organizations do not know how to go about the task of conceptualizing, developing, installing, and effectively using a maintenance management information system. Consequently, many have hedged their primary responsibility of specifying what information they want, how they will get it, and how they will use it effectively. Into this void have jumped many who claim they can help. But some may not understand the special information needs of maintenance in the mining industry. By default, the whole scene of maintenance information systems has become confused and in need of clarification. A Starting Point Four broad phases are used to organize the steps by which maintenance acquires an effective maintenance management information system and initiates its use. Conceptualization is an identification of needs, factors that influence development, and an assessment of support for the system. Development is a plan of action for bringing the conceptualized information system to life. Consideration is given to the practical aspects of collecting field data, processing it, and converting it into useful information. Implementation consists of those actions by which the newly developed system is integrated into the organization and put in the hands of its personnel so that its value can be derived. Use addresses the means by which personnel use the information to accomplish maintenance improvement objectives. These phases are preceded by establishing the need for the information system. They are followed by a requirement to measure its success (Fig. 1). Conceptualizing the Information System Responsibilities - The information system user must acquire the system best suited for its use. A step-by-step process is essential to avoid many serious problems than can easily occur. The most important step is to clearly identify its objectives and lay out a course of action to achieve those objectives. The major reason for the failure to achieve information objectives is not carrying out this responsibility. Maintenance cannot delegate the task to data processing. It can obtain technical assistance but it must retain responsibility. Nor can maintenance purchase

Jan 7, 1985

By M. P. Adamson, Bryson D. Trexler, R. V. Ramani, Mark L. Koenig, V. Rajaram, Alan Burton, C. S. Crocker, Kim. Y. C., A. G. Law, William C. Larson

Another record year vas experienced by the mining industry as the value of nonfuel minerals production soared to $23.5 billion-up 15% from the previous year. Based on preliminary estimates by the Bureau of Mines, US mine output was the basis for about $225 billion worth of minerals-derived materials. Metal mining accounted for $8.5 billion, or 36%, of the total. Of the 22 metals produced in the US, 11 showed quantity increases, while 16 had value increases. Business Week's survey of 1200 companies (Mar. 17, 1980, pg. 81) revealed that 21 metals and mining companies had sales increases of 30% with earnings up 104%. Higher prices and increased demand for metals were primary factors behind the improved performance. The steel sector reported a 30% decrease in earnings for the year, but the slump was primarily the result of a $383-million loss for US Steel coupled with a 17% decrease in earnings at Inland Steel. Other firms did well in 1979, with earnings up 279% at Interlake and 269% at Kaiser Steel. Major obstacles facing the steel industry included substantial tonnages of dumped and subsidized imports, environmental expenditures, runaway inflation, and price controls. Turning to nonmetallic production, of the 49 commodities reported by the US Bureau of Mines, 29 had higher outputs and 39 had increased values. While production statistics may seem encouraging, government regulations continued to stifle industry efforts to develop needed new capacity. A case in point may be publication of the Environmental Protection Agency's simplified procedure to streamline five permit programs by consolidating them into one. The five included: hazardous waste management, national pollution discharge elimination, dredge and fill, underground injection control, and prevention of significant deterioration. The new procedure, published by EPA in the Federal Register, took 149 pages. AMC's J.A. Overton, Jr., noted that (MCJ, Nov. 1979, p. 39) to visually depict the complexity of the proposed regulations, Sell Oil Co. developed a flow chart mapping a path through one section. The chart required 10 days to map and two weeks for two draftsmen, using computerized drafting techniques, to plot the results. Overton said, "Perhaps the greatest service EPA has performed in preparing its proposal is to demonstrate again that federal regulations have become a complexity compounded by confu¬sion that culminates in a conundrum." While regulatory agencies promulgate new regulations, the lack of a cohesive government minerals policy remained a matter of significant importance to mine management. A non-fuel minerals study to provide input to the Administration for a national minerals policy program was a dismal shock, according to Rep. Jim D. Santini (D-NV). He said the first phase of the program, which was to identify industry problems, failed to come to grips with the hard decisions necessary regarding these problems, their impact, and their interrelationship. Santini said the report failed to acknowledge government's role in the process and absolved the government of any adverse influence on the domestic minerals industry. In response to the phase one report, eight senior mining executives testified before Santini's House Subcommittee on Mines and Mining concerning mineral supply, access to foreign sources, environmental protection, conservation and recycling, national security, and capital formation. By year's end, Rep. Santini promised to prepare a hard-hitting committee report on problems affecting the non-fuel minerals industry to provide a better base for subsequent policy recommendations (MCJ, Jan. 1980, p. 34). On the technological front, interest again centered on mechanization, automation, and scaleup to improve and expand production. Computer use increased for calculation of ore reserves, mine planning, operations research, and mineral processing.

Jan 5, 1980

By P. M. Soupios

An integrated suite of environmental methods was used to characterize the hydrogeological, geological and tectonic regimes of the study area, to determine the geometry of the landfill (depth and spatial extent of electrically conductive anomalies), to define the anisotropy caused by bedrock fabric fractures and to locate potential zones of electrically conductive contamination. The study area which covers an area of about 0.08 Km2 is a municipal solid waste site (MSW), located 20Km West of Herakleion in the central part of Crete (S. Greece). The landfill is located in a northwest oriented valley with highlands to the northeast and southwest. The combination of geophysical methods and chemical analysis are commonly implemented for characterization and management of land-fills. At the particular site, geophysical methods were focused on studying the electrical insolating properties of the materials, electrical con-ducting properties of fluids (leachates) and electromagnetic properties of the disturbed regolith materials in the burial zones. Three different types of geophysical surveys were performed: 1) 2D electrical tomography (ERT), 2) electro-magnetic measurements using very low frequencies (VLF) and 3) electromagnetic conductivity (EM31). Furthermore, in order to calibrate the geophysical methods dielectric spectrometry measurements are conducted. To evaluate the environmental degradation of the area, soil and liquids samples were collected.

Jan 1, 2005

By G. I. Voiloshnikov, V. E. Dementyev

Wide use of adsorption methods for metal recovery is one of the most important achievements of gold metallurgy of the last decades. Anion exchange resins or activated carbons may be used as adsorbents. This paper presents a brief overview of IRGIREDMET effotts on development, testing and implementation of techno­logical flowsheets using both adsorbent types and attempt to compare their technological properties. IRGIREDMET Institute was founded in 1871 as a laboratory for gold smelting and gold ore testing. In 1930 it obtained the status of the Research Institute. IRGIREDMET functioned as leading research center for gold mining industry in the former Soviet Union. At present it is the Open Joint Stock Company dealing with all s01t of research and design works for gold min­ing (ranging from geology to gold refinery). IRGIREDMET started development of ion exchange technol­ogy in 19641• The first stage of work (1964-1970) including development of principle flowsheet of RIP process was carried out jointly with the Moscow VNIIKHT Institute (All-Russian Scientific Research Institute of Chemical Technology)2• The flowsheet of RIP process is presented in Fig. I and includes the following basic operations: gold adsorption from pulp; washing of loaded resin; acid treatment; thiourea adsorp­tion; gold elution by acid thiourea solution; resin washing to remove thiourea; alkaline treatment of resin; resin washing and gold recovery from eluates by electrowinning. The technology developed and equipment used were tested for treatment of clayey ore of Kuranakh deposit at a pilot plant of the Lebedinsky gold complex ("Aldanzoloto" company, Yakutia, Russia). The pilot plant throughput was 200 t/day. At the same time IRGIREDMET conducted the tests on ion exchange technology for ore and concentration products of oper­ating gold plants and for ores of new deposits. Laboratory and pilot plant tests were done and developed initial data and process design criteria for designing of 15 plans of the former Soviet Union. In the course of this work a great number of newly synthe­sized anion exchangers including strong and weak base resins was studied, the peculiarities of anion exchange resins using two different raw material types ( ores and concentrates includ­ing clayey, carbonaceous, silver bearing, arsenic, antimony and base metals bearing products); some theoretical regularities of adsorption process were investigated; the technology of two stage adsorption of gold and silver from silver gold ore, improved flowsheet for gold elution and resin regeneration, design of adsorption equipment including pulsation columns, design of electrowinning cells were developed; conditioning technology for pulp with complex ion composition prior to gold and silver adsorption was developed, etc.3·8• IRGIREDMET specialists took part in commissioning of practically all the plants in the former Soviet Union using RIP and RIL process including Lebedinsky, Kuranakh, Sovetsky, Karamken, Matrosov's plant and Mnogovershinny plants. Due to this works RIP process became standard technology for Russia and the CIS.

Jan 1, 1998

By Mahmut Altiner

Magnesium is listed as a key critical raw material in a report by the European Commission that evaluated raw materials based on supply risk versus economic importance [1]. It has extraordinary properties, such as nontoxicity, inertness, high melting point, stability, low weight, chemical resistance, electrical insulation, optical transparency, corrosion-resistant behavior and refractory nature. It is widely used in the areas of agriculture; the environment, for antibacterial and water-purification purposes; electronics; automotive; drug delivery; ceramics; construction, as filler in brick; and concrete production, in the forms of magnesium oxychloride and magnesium oxysulfate. Magnesium oxide (MgO) particles can be produced through the calcination of magnesite ore at different calcination conditions depending on the desired quality of the product, classified as fused (2,800 °C), dead-burned (>1,400 °C), hard-burned (1,000–1,400 °C), or light-burned MgO (700–1,000 °C) [2]. However, the global availability of magnesite ore is limited.

By R. H. Lesernann

Copper has held up extremely well as we have moved through the trough of the economic cycle. In the aluminum, lead and zinc industries at least half of the Western World production was unable to cover its operating costs at the prices prevailing throughout most of 1991. In copper, less than 10 percent of production ran in the red last year. Most of that was in Peru where the government has not allowed its currency to devalue at the same pace as inflation. Thus, mining company revenues in local currency terms have risen very modestly while costs have escalated tremendously. Excluding such governmentally imposed distortions, virtually every mine in the Western World ran at a profit last year. While copper producers have to be extremely happy about staying in the black through the trough of the consumption cycle, consumers have to be worried. They must be asking themselves, "If copper held at over $1.00 in a year when consumption declined, how high will prices go when demand picks up?" Consumers are accustomed to paying high prices at the peak of the cycle, but they are also accustomed to paying low prices at the trough of the cycle. What went wrong this time?

Jan 1, 1992

By E. D. Dainty, B. E. Enga, D. L. McKinnon

Following considerable laboratory studies concerning particulate control systems for underground use, a program was commenced to develop practical real life implementation systems. This program covers the implementation in seven mines of particulate filter systems to 10 engine models from three engine manufacturers in 12 different vehicle types, a total of 24 vehicles. As the program ran, one mine had to drop out and a mutually agreed upon system could not be arrived at for another. Hence, the program included five Canadian mining companies. Several of the vehicles trialed at Brunswick Mining were trialed by Brunswick themselves. Nonetheless, the data is presented. This paper covers the work performed from the beginning to the end of the demonstration program. Data collected to properly select and design filter systems along with technical conclusions of the demonstration are presented. Also presented are the results of a survey that highlights the participating mines' acceptance of the technology.

Jan 1, 1991

By Hans Smit

The face of mining as we know it is set to change in the very near future as mining companies all over the globe realize the implications of accessing the vast mineral wealth contained on, and under, our ocean floors. With some land based mineral deposits being fully exploited across the world, deep sea mining is set to revolutionize global approaches to the search for precious metals and other sought after minerals. [ ] Previously, the financial implications of deep sea mining outweighed the benefits of access to the mineral rich deposits identified all around the world. The process of marine diamond mining off the west coast of southern Africa started in the late 1950?s by a Texan, Sammy Collins, and subsequently continued by De Beers to the present day, has proven to be an excellent proving ground for the development of technology and mining systems that ensure a sustainable and financially viable business. [ ]

Jan 1, 2008

Allis Mineral Systems' Grate- Kiln system can economically convert low-grade iron ore into high- grade iron-ore pellets suitable as feed material for blast furnaces for making steel. These pellets can also be fed to a direct reduction process for making sponge iron. The Grate- Car system combines the company's [ ] Grate-Kiln system for iron-ore pelletizing and Accar system for iron- ore reduction. The Grate-Car system consists of a traveling grate, a ported rotary kiln and a rotary cooler that are coupled in series. This system eliminates the pelletizing plant's rotary kiln and annular cooler. It reduces the need for some auxiliary equipment. This results in lower capital cost and lower energy consumption. The traveling grate provides that static bed handling necessary for heat hardening pellets before their discharge into the kiln. The ported kiln provides the reducing atmosphere with a more precise means of controlling the kiln axial temperature profile than conventional rotary kilns.

Jan 1, 1997

By C. D. Iles

The philosophy used in designing and writing Duval's Sierrita mine design programs is discussed. The Sierrita Incremental Design system (SID) is reviewed with emphasis on Pushback Extraction Design (PED). FED is a method by which predesigned pushbacks are sequenced into an optimized period-by-period design that can be directly utilized in daily operations as well as intermediate and long range planning. The advantage of SID is that it provides a practical, mineable open pit design. PED is not a theoretical design system; it is in operational use.

Jan 1, 1981

By J. L. Kohler

Control circuit failures contribute significantly to continuous miner downtime. The time required to put the machine back into production is a function of the time required to diagnose the problem, obtain parts, and make the repair; of these, the diagnostic time is often the largest. A microprocessor-based system with the ability to diagnose all of the common control circuit failures was designed, installed on a continuous miner, and tested. This paper summarizes the development of the diagnostic system, and provides examples of the system's diagnostic capabilities. Implementation issues and the hardware used for the prototype are briefly described. Finally, comments on the system's performance are presented.

Jan 1, 1993

By Robert J. Neubauer

I was asked here today to discuss the first long-term freight rate contract ever signed by a U.S. railroad and shipper. That was ICG's contract with the hoosier energy company to haul coal from southern Illinois to an Indiana power plant. I?ll also talk about why you can expect to see more such contracts in the future But first, let me tell you a little bit about the Illinois central gulf railroad and why coal contracts are so important to us. ICG is a principal mid-American freight carrier. We are one of the ten largest railroads in the united states and one of the ten largest coal-hauling railroads. ICG links the industrial great lakes region with the warm weather ports of the gulf coast. Some of the major points on our route map are Chicago; Indianapolis; Omaha; Kansas City; Memphis; St. Louis; and, of course, baton rouge and new Orleans.

Jan 1, 1981

By Juan L. Yarmuch, Gonzalo Sepulveda

Medium- and short-term mine planning holds immense significance in optimizing production and resource utilization. It serves as a vital link between the long-term strategic plans and the day-to-day operational activities of a mine. In this context, a mathematical formulation is proposed to incorporate mining width constraints into the production scheduling process, aiming to maximize the net present value (NPV) of the schedule while ensuring enough room for the operation of mining equipment.

Jul 1, 2024

By J. W. Furlan

Due to the time consuming nature of froth testing in a laboratory environment, it is required, due to practical limitations, to develop a method of applying a single set of performance test data (Visintainer & Whitlock, 2012) obtained for a single froth pump at a given speed, to a range of froth pump speeds and sizes, % air contents, and liquid viscosities. Head, efficiency, and flow all vary as functions of both liquid viscosity and air content. As such, it is necessary to develop a technique which can be used to interpolate from existing test data in order to predict froth pump performance for pumps of various sizes running at variable speeds. Details of the technique developed in order to achieve this goal as well as the insights gained from analysis of the resulting regressions are discussed. Additionally, comparisons are made between the performance predictions of this test-derived method and those of the Hydraulic Institute Standard for the effects of liquid viscosity on centrifugal pump performance (ANSI/HI 9.6.7-2010).

Feb 27, 2013

By Wade W. Bristol

The Jerritt Canyon Joint Venture project is located about 80 km (50 miles) north of Elko, NV. It produces approximately 9.3 t/a (300,000 oz/year) of gold. The joint venture is 70% owned by Independence Mining and 30% by Meridian Gold. Independence Mining is the operator. The Jerritt Canyon Joint Venture started operations in 1980 by mining a series of open-pit mines. The discovery of deeper ores during early 1990s made it neccessary for the operation to mine a large portion of its annual production by underground mining methods. Underground production started in 1993 at the West Generator Mine. This was followed by the Murray Mine in late 1994. Two additional underground projects are being developed. Underground mining at Jerritt Canyon is scheduled to produce about 3.9 t (128,000 oz) in 1997.

Jan 1, 1997

By Thiemo Hagedorn, Gerhard Weissbach, Ralph Bergmann, Wilhelm J. Kogelmann

The objective of this unique method is to multiply the advance rate of conventional tunneling while improving both ground control and the safety of the crew. The high advance rate is achieved by concurrent rock excavation and installation of concrete lining at fully synchronized work phases. The novel TACHUSMATIC system consists of a high-production, robotic cutter boom, an automated slip form and continuous mucking and conveying. This novel system utilizes the inherent bearing capacity of the rock that is activated by a small, initial deformation. The supporting resistance is then provided both by the rock mass itself and the pumped in concrete lining. During the hardening of the concrete, the hydraulic cylinders of the TACHUSMATIC automatically apply the appropriate support resistance, thus providing safe ground control. This novel tunneling system is practically a fully automated, fail-safe NATM that does not require highly skilled engineers and crew. The TACHUSMATIC can be used for construction of circular, rectangular and horseshoe-shaped tunnels. When equipped with the Alpine Multi-Tool Miner (MTM) systems, the machine can work in a wide range of ground and rock conditions. The use of novel, diamond-tipped cutter bits (picks) will increase hard rock cutting capabilities and will significantly reduce pick consumption. Because of its moderate cost, the novel system is an economic alternative to conventional tunneling even at short drivages. A new installation is cost-effective for tunnels longer than 750 meters (2,500 ft.) At a typical tunneling project, the cost permeter is about 25% lower than with conventional methods.

Jan 1, 2003

By R. D. H. Bedd

GEOLOGICAL NOTES The area under discussion is situated approximately 150 m north of Loraine No. 3 shaft at ± 1500 m below surface. The block being mined extends from about halfway between 50 and 52 levels to just below 48 level (see Fig. 1). The EB reefs form the lower portion of the EB zone of the Elsburg series and in the 81 line area the various conglomerate bands show the best development. Up to 18 separate reef bands are developed and the distance from the bottom conglomerate band to the top is ± 26 m. The thickness of the quartzite which forms the middling between the various reef bands varies from 0.1 to 2 m. The strike is roughly north-¬south. The reef bands are payable individually. However, owing to the steepness of the reefs and their close prox¬imity, conventional mining would prove very difficult. The nonpayable reef bands all carry varying amounts of gold and the whole area is a massive payable reef body approximately 26 m wide and 60 m high at 300 to 400 m on strike. The major structure of the area consists of a north¬-south trending syncline. The western flank of the syncline has a steep easterly dip, in places approaching the vertical. This flank cuts out against the overlying rocks of the EA zone. The latter forms a noncon¬formity with the underlying EB zone. The eastern flank of the syncline dips to the west at an average dip of 0.17 rad (10°). The 0.1 rad (6°) north-plunging syncline was sub¬jected to severe faulting. The most important are the Nos. 4 and 10 faults, both of which have a roughly north-south strike. The No. 4 fault dips to the west at an angle of 1.04 rad (60°) with a downthrow of approximately 150 m to the west. The dip of the No. 10 fault is 0.69 to 0.81 rad (40° to 50°) west, with the western block having moved 150 m updip relative to the eastern block. The EB reefs are being mined by the sublevel stoping method down from the sub-outcrop position to the cutoff against the No. 4 fault. On the downthrow side of the No. 10 fault, that is just above the 54 level, two reef bands are payable and will be mined conven¬tionally. ROCK MECHANICS From the section in Fig. 1, it can be seen that the area under consideration is surrounded by stoped-out areas. This will naturally change the stress distribution extensively. The only way to calculate the stress field for such a geometry is to use the finite element method. It was established that this method worked for the present geometry and so the method was applied to predict the stress behavior of the final geometry. From the results, it appeared that the effect of the old stopes was to destress the area. There are, however, some areas-such as the 50 and 48 north footwall haulages-which will require continuous attention. In these two footwall haulages, deterioration can take place due to tensile stress which

Jan 1, 1982

By H. Keith Henson

The purpose of this paper is to describe the possible application of two relatively new technologies: flat load cells, and improvement seeking computer programs (referred to in the literature as heuristic programs); to a very old problem, inventors control of ball mills. The combination of these two technologies seems most applicable to grate discharge mills operated closed circuit. It is possible that other applications such as controlling the inventory of autogeneous mills might also be found. Direct control of ball mill inventory has been the goal of a number of investigators over many years. Many patents and articles on the subject have claimed success, but operating installations are hard to locate. An excellent review of the advantages and problems of direct inventory control is contained in an article by Oswald and Ziegler (1). Figure 1 is a conventional representation of throughput vs. inventory with other factors such as slurry consistency and rock twee held constant. There are common sense, physical arguments? and experimental evidence supporting the existence of a unimodel curve of this twee? though direct measurement of the shape is difficult and how it moves as rock type changes is riot well known. Figure 2 represents the problem faced by all mass measurements of inventory, a tare weight of about 25 time the inventory mass. Worse yet, the tare is variable due to ball and liner wear and the Periodic addition of balls. The problem then breaks down into two parts: 1) obtaining enough mass

Jan 1, 1977

By François Martin, Roland Plassart, François Laigle

"MOTIVATION The storage of nuclear waste gives rise to a lot of innovative issues for underground space technologies. The nuclear activity of waste generates temperature effects on both the lining and the surrounding rock mass. In order to avoid these unfavorable effects, the number of waste containers is controlled in the gallery section. The galleries section has consequently ranged from a diameter of 4 to 6 meters. However, in the case of intermediate-level nuclear waste, due to the lower temperature, the number of containers in a cross section of a gallery is assumed to be safely increased. Despite technical difficulties, a large diameter gallery (typically 12m excavated) seems to be more cost effective. This paper demonstrate the feasibility of such a gallery in the specific context of CIGEO (French Geological Repository Project for nuclear waste), located in the Bure area in the north-eastern of France. The geology of Bure at the project’s depth (about 520 meters) consists in a quite homogeneous layer of Callovo-Oxfordian argillites (COx argillites). With such an overburden, this rock is known to have a high plastic behavior with significant time-dependent effects. These mechanical properties require complex behavior laws in computational modeling used for the design of underground facilities.EXPERIENCE OF LARGE DIAMETER GALLERIES In order to identify some specificities in the design of such openings, a research has been run on existing galleries of large diameter (more than 10 meters), like the tunnel of Chamoise, Lötschberg, Tartaiguille or Saint-Martin-La-Porte. Two of these representative examples are detailed hereafter. Tunnel of Chamoise (French Jura) In spite of a lighter overburden (400m versus 520m), an anisotropic initial stress state and a variable rate of carbonates, the highway tunnel of Chamoise is quite similar to the planned gallery for intermediate-level nuclear waste, with a geology very close to the CIGEO’s one in some sections. Because the northern tube encountered difficulties during the excavation a few years before (several over-excavations and spalling, especially in the bench), the excavation of the southern tube has been carried out in two steps, but with a high bench and a small curved invert (Figure 1). The design of the immediate support consists in long bolts and a shotcrete layer. The concrete lining is about 50cm thick."

Jan 1, 2016

By R. L. King

An expert system is a computer system that performs tasks that are considered by most people to require human expertise. Since their functioning relies heavily on large bodies of knowledge, they are often referred to as knowledge-based systems. These systems can contain the skill and knowledge of an individual human expert or the combined knowledge of a group of experts. The knowledge embodied in an expert system can consist of specific facts, HSHA regulations, heuristics or rules-of-thumb used by the expert, and databases. The purpose of this paper will be to introduce the participants in the ?Increase Competitiveness Thru Innovation? session to expert systems. This introduction will enable the participants to better appreciate the balance of the papers to be presented in this session. These papers will deal with specific mining problems and a proposed innovative solution. This paper will discuss such topics as problem-solving paradigms, rules of inference, knowledge representation, and the structure of expert systems.

Jan 1, 1989