Search Documents

By Thomas R. Kuesel

Just south of the L'Enfant Station in Seventh Street, the WMATA system splits into two branches, each comprising twin single-track soft ground tunnels. Immediately at the beginning of the section, Interstate Route 95 crosses transversely over the alignment in open cut, removing most of the cover over the tunnels. Seventh Street spans I-95 on a bridge whose abutments and pier also cross transversely over the four tunnels, with their footings as little as 7 feet above the tunnel crown (Figure 1). Soil investigations revealed that the tunnels would have to be driven through mixed face, with a clay invert and a waterlogged sand crown. Thus, there was a good chance that the roof would fall in, leaving 1-95 and the Seventh Street Bridge in a crater. Since there was no roof that could be counted on, these were figuratively (and potentially truly) Washington Metro's Topless Tunnels (Figures 2 and 3). DESIGN ALTERNATIVES Faced with this situation, the designers considered several options, including temporary detours that would permit demolition of the bridge and building tunnels under the 1-95 Freeway by cut-and-cover methods. They also investigated structural underpinning of the bridge and freeway pavement, and strengthening the bridge abutments and center pier by post-tensioning to enable them to span over any subsidence trough.

Jan 1, 1976

By William I. Weisman

Natural sodium sulfate is obtained from both brines and solid deposits. The brines almost always contain other salts with sodium chloride predominating. They may be obtained from surface occurrences, such as Great Salt Lake or from shallow subterranean deposits. Althought many sodium sulfate minerals are known to exist, only three are of commercial significance. These are thenardite, the anhydrous salt; mirabilite, the decahydrate; and glauberite, the double salt of sodium and calcium sulfate. The reader is referred to Weisman (1983) for details on mineralogy and geology of sodium sulfate minerals. Sodium sulfate deposits, either in the form of brine or solid minerals, are found in many parts of the world. Natural sources, by far, provide the primary supply of commercial grades in Canada, Mexico, Spain, and probably in the U.S.S.R., Chile, Argentina, and Turkey. In the United States, production of natural material and synthetic (including by-product) are about equal in volume and have been for many years.

Jan 1, 1984

By S. R. Maynard

The Ortiz Mountains lie 25 mi (40 km) northeast of Albuquerque, New Mexico in Santa Fe County (Figure 1). Approximately 350,000 ounces (11,000 kgs) of gold have been produced from the Ortiz Mountains since the 1820s. The 100 square mile (260 square km) Ortiz Mine Grant was secured from the Mexican government in 1833 by Jose Francisco Ortiz for the purpose of mining gold. The Grant has remained more or less intact to the present. The mineral rights are currently owned by Ortiz Mines, Inc. of Joplin, Missouri and are leased to LAC Minerals, USA, Inc. (west half) and to Gold Fields Mining Corporation (east half).

Jan 1, 1989

By Frank Ruskey

Many problem associated with potential hazards in a proposed mining area may not be adequately appraised because of the complexity of trying to evaluate them in terms of production requirements. Frequently, encountered hazards have to be dealt with by day-to-day operational schemes that require the miner to either force a way through them or avoid them by blocking off the area in which the hazards are known to exist. These considerations, after the mine is in full production, can be costly to the mine operator and can increase the danger to the miners. A hazard-evasion computer program for mine planning that will appraise proposed mine layouts on the basis of existing or potential hazards has therefore been developed. The program is intended for use during the planning phase of a mine, where it should assist the mining engineer in making decisions in mine design. With the output from this program, the mining engineer should be able either to corroborate his judgments made on the basis of mining experience or to find unanticipated solutions to design problem. The program tests hundreds of possible solutions to a mining layout, and by sorting out the 10 best (based on a criteria of minimum-abandoned-coal) limits the number of choices that need to be considered and appraised for applicability to a given mine scheme.

Jan 1, 1973

By James R. Jones

The purpose of these opening remarks is to establish a viewpoint, which will form part of the basis for later discussion between panel members and between the panel and the audience. Important issues will be identified. These ideas in the form of questions and answers, hopefully, will provide the necessary background for the ensuing dialogue. Is There An Urgency? We do not believe there is the urgency for immediate control legislation that has been depicted by some Congressional sources, environmental groups, and certain segments of the press. Emissions of sulfur oxides in the United States have decreased 15% during the 19701s, even though the use of coal by electrical utility power plants increased nearly 80%. That trend will continue as the older plants are phased out of operation and are replaced with newer plants, which must meet the New Source Performance Standards (NSPS) requiring the use of "scrubbers" to reduce sulfur dioxide emissions by 70% or 90%. The present Clean Air Act Amendments of 1977, by allowing the logical progression of economic events- -depreciation, retirement, replacement, and growth--will produce approximately a 50 % reduction in SOX emissions by the year 2015.

Jan 1, 1984

By C. A. Carroll

In February of 1989 the Mission Complex of ASARCO Incorporated completed an expansion of its concentrator, raising the base mining rate from 25,900 TPO (28,500 STPD) to 37,200 TPO (41,000 STPD). The expansion was accomplished by extending the twelve existing Mission ball mills in length, by installing the grinding circuit from Asarco's idled Sacaton Concentrator along with various auxiliary equipment, and by modifications to the flotation circuit and water systems. This paper will give a brief history of the consolidation of the ore body which made the expansion possible and explain the various parts of the expansion.

Jan 1, 1990

By W. C. Dailey

The Groveland Mine. Concentrator and Pellet Plant are located in Dickinson County in the Upper Peninsula of Michigan. The construction of the Groveland Mine in 1959 was the first United States operation for the Hanna Mining Company in mining and beneficiating low grade iron formations. (1) The plant produced concentrates at 675,000 lt/yr. (746,600 mt/yr.) which was used as sinter feed in the steel industry from 1959 to 1963. Upon completion of a pelletizing plant in April, 1963 production increased to 1.6 million lt/yr (1.77 million mt/yr) of pellets. After the plant expansion, mill capacity increased to 2.2 million 1t/yr (2.43 million mt/yr.), producing 2.0 million It/yr (2.23. million mt/yr.) of pellets assaying 63.0% iron and 7.0% silica. (1) Prior to April, 1971 the concentrator operated with a flowsheet consisting of spirals and a petroleum sultonate iron notation circuit. This was presented in a paper by R. R. Smith and E. E. Sougstad, "Flotation at the Groveland Mine" in the 50th Anniversary Volume entitled "Froth Flotation". The ore body at the Groveland Mine consists basically of two types of iron formation (34% total iron, 20% magnetic iron head), the most abundant of which is a bedded magnetite-chert-silicate rock. This is intermingled to varying degrees with non-bedded to poorly bedded hematite-clastic quartz-chert rock. (1) In developing the Groveland flowsheet, Humphrey spirals were used because of their ability to produce a coarse size concentrate at a low cost. Tests run in the lab and pilot plant on the Groveland ore showed a. spiral weight recovery in the 27 to 33% range. (2) However as the ore body was opened up, it became evident that the most abundant ore type did not liberate at the coarse size fraction needed for spirals. Plant recoveries for the first five years operating the spirals only averaged from 10 to 15% weight recovery. In 1971 the weight recovery on the spirals had dropped to a low of 5%. To increase iron recovery and at the same time increase plant capacity with the least effect on the present operations the mill flaw-sheet proceeded from a spiral-notation circuit to an all flotation circuit to a magnetic separator-flotation circuit. During this changeover the concentrate grade also had to be improved to meet market demands. To do this the following changes evolved

Jan 1, 1978

By Peter N. Blakey

The Kidd Creek orebody is a massive, base metal deposit with widths up to 600 feet and a strike length of 2,200 feet. The orebody dips eastwards at 70° to 80° and is amenable to sub-level open stoping. Five distinct zones (Fig.1) based on structural wall rocks and ore type parameters occur within the orebody. These zones are: Zone 1 - Mainly copper stringer ore Zone 2 - Massive pyrite ore Zone 3 - Massive sphalerite and chalcopyrite Zones 4 and 5 - Are banded and variable in nature Compressive strengths vary from zone to zone with an average of 22,000 p.s.i. and highs of 40,000 p.s.i. The mine is developed from a ramp extending from surface to the 1600 foot level and a 24 foot diameter, 3050 foot vertical shaft with main levels at 400 foot intervals. Sub-levels at 100 foot vertical intervals are driven from the ramp. transverse and longitudinal pillars to support the crown pillar left under the open pit. The layout allows an opportunity to enlarge or reduce pillar sizes and also to introduce fill on any horizon should conditions warrant. It also offers a large number of operating stopes spread through the various ore zones giving good ore blending characteristics. All ore from the stopes is handled through drawpoints using Wagner ST8 scooptrams with an average one way haul length of 650 feet.

Jan 1, 1976

By Mahmut Altiner

In the present study, the use of horse residue as an alternative reducing agent in the roasting of a low-grade goethite ore containing 17.60% Fe was investigated. Additional roasting experiments were carried out with corn starch and without reducing agent for comparison. Each roasted product was characterized using X-ray diffraction (XRD) and magnetic hysteresis (MH) loops analyses. The optimum roasting conditions (roasting temperature [400, 500, and 600 °C], reducing agent/ore ratio [0.04, 0.08, and 0.12], and roasting time [10, 20, and 30 min]) were investigated based on the Taguchi approach. The magnetic susceptibilities of the roasted sample with the addition of reducing agents were higher compared to that of the roasted sample in the absence of reducing agent. Experimental findings indicate that the horse residue can be used as the reducing agent. Finally, an iron ore concentrate assaying a Fe grade of 62.12% with a recovery of 56.93% was prepared using wet magnetic separation at a magnetic intensity of 0.5 T from the ore roasted under the following roasting conditions: roasting temperature of 500 °C, reducing agent/ sample ratio of 0.08, and roasting time of 30 min.

Jun 3, 2020

By R. J. Sweigard

Productive and beneficial use of reclaimed surface-mined land is a concern of industry, government, and the general public. An investigation was conducted of current mined-land planning practices to aid in the development of a procedure for evaluating postmining land use potentials. This investigation considered both the industry and public planning approach to mined-land planning and was based primarily upon three case studies conducted in various regions of the United States. The site planning process that was developed during the investigation is iterative in nature and has been designed specifically for the mining industry. It centers around preliminary studies and data collection, development of land use alternatives, and the evaluation of alternatives. Use of this process facilitates the evaluation of economic, environ- mental, and social impacts of postmining land use alternatives. In this paper, the process is presented along with a description of its development, application, and significance.

Jan 1, 1984

By W. M. Ma, Daniel W. H. Su, K. Centofanti, Yi Luo, W. L. Zhong, Syd S. Peng

2.1 INTRODUCTION When total extraction of an opening of sufficient size is reached in a horizontal coal seam, the roof strata in the overburden deform continuously to reach a new equilibrium condition. The severity of deformation decreases upward toward the surface. As the downward saggings of the strata propagate and reach the surface, there will be a depression zone on the surface directly above, but extending beyond the edges of the underground opening. This is the surface subsidence basin or surface subsidence trough. The surface subsidence basin is circular in plan view, if the coal seam is horizontal and the mined-out opening is square in shape. But it is rectangular with rounded corners or elliptical if the coal seam is horizontal and the mined-out opening is a long- and thin rectangle or a short-rectangular, respectively (Fig. 2.1). Most underground openings (e.g., longwall panel) assume rectangular shape when total extraction has been completed. Theoretically the edges of the subsidence basin are the points of zero subsidence. But it is difficult to exactly locate the points of zero subsidence. Therefore in practice the points with vertical subsidence of 0.4 in. (10 mm) are used. The final subsidence basin is that which exists long after the mining has been completed, because its magnitude and shape are quite different from the dynamic subsidence basin formed while the face is moving. 2.2 CHARACTERISTICS AND TYPES OF DEFORMATION IN THE FINAL SUBSIDENCE BASIN For a horizontal coal seam, every point in the subsidence basin moves toward the center of the basin. Subsidence is maximum at the center of the basin. Any cross-section that passes through the point of maximum subsidence and either parallel to AB or CD line (Fig. 2.1) is a major cross-section along which principal directions of surface movements occur. However among those infinite numbers of major cross-sections, two specific ones are of special significance, not only because the magnitudes of surface movements are the largest, but also because they are the most easily identifiable directions, i.e., one that is parallel to the faceline at the center of the basin (CD in Fig. 2.1) and' the other is that perpendicular to the faceline but parallel to the diction of face advance (AB in Fig. 2.1). Nearly all the subsidence data obtained in the US have been derived from these two cross-sections, although some cross- sections parallel to CD but near the edges of the panel have also been included. In addition to moving horizontally toward the center of the basin, every point in the basin also subsides vertically. The magnitude of subsidence increases toward the center of the basin. Therefore surface subsidence is a three-dimensional problem and should be treated so in all cases. On all the major cross-sections, only principal subsidence and principal displacement occur. Since subsidence and displacement vary continuously in every major cross-section, three additional deformation components are de- rived, i.e., slope, curvature, and strain. On all other non-major cross-sections on the other hand the five components are accompanied by two additional components, i.e., twisting and shear strain. The seven components of the surface movement are defined as follows (Fig. 2.2): 1. Subsidence, S. On any cross-section, the vertical component of the surface movement vector is called surface subsidence. It generally points downward. But sometimes it points upward in areas ahead of the faceline or beyond the edges of the opening. In such cases it is a surface heave which is usually less than 6 in. 2. Displacement, U. On any cross-section, the horizontal component of the surface movement vector is called surface horizontal displacement. It generally points toward the center of the subsidence basin. But in steep terrain, it moves along the downdip direction 3. Slope, i. On any cross-section, the difference in surface subsidence between the two end points of a line section divided by the horizontal distance between the two points is called the surface slope of the section. 4. Curvature, K. On any cross-section, the difference in surface slope between two adjacent line sections divided by the average length of the two line sections is called the surface curvature of those two line sections. There are two types of curvature: con- vex or positive curvature and concave or negative curvature. 5. Horizontal strain, e. On any cross-section, the difference in horizontal displacement between any two points divided by the distance between the two points is called horizontal strain. If the distance between the two points is lengthening, it is tensile strain with positive sign. Conversely, if it is shortening, it is compressive strain with negative sign 6. Twisting, T. On the surface of the subsidence basin, the difference in slope between two parallel line sections divided by the distance between the two line sections is called twisting. 7. Shear strain, y. Shear strain is the changes in internal angles of a square on the surface of the subsidence basin or on any major cross-section. It is the summation of the differences in incremental (or decremental) lengths between the two opposite sides divided by the original distance between the two opposite sides. More precisely, the surface deformation indices (i.e., slope, strain, curvature, twisting and shear) are defined by derivatives of surface movement components. For simplicity, the x- and y-axes of the cartesian coordinate system are set to be parallel and perpendicular to the cross-section of interest, respectively. In such a coordinate system, slope and curvature along x direction are the first and the second derivatives of the vertical components (S) of surface movement with respect to x, respectively, or i, = ds/dx and kx = d2s/dx2. Horizontal strain along x direction is the first derivative of the component along x direction of the horizontal displacement,

Jan 1, 1992

By J. Pancher

"Stainless steel castings are used in abundance in the mineral processing industry. In rotary kiln-based processes, these castings are typically exposed to high loads, corrosive elements, mechanical abrasion and cyclical temperature exposure. In the processing of iron ore pellets, it is typical to use a traveling grate preheater or a straight grate furnace to heat the pellets. Both machines have a conveyor that is mainly comprised of high temperature stainless steel castings. It is these castings which are the focus of this study. The expected life of these castings can be quite short, their cost quite high and the effort to replace them very labor intensive. The goal of this study was to gain a better understanding of the failure modes of these castings. With that basic understanding, modifications to the alloy content could be made to achieve a new operational paradigm. That paradigm could be a cheaper casting, a longer lasting casting or a casting tailored to special operating conditions (for example, highly corrosive gases or sudden temperature shocks). Over the past 4 years, Metso Minerals has sought to achieve a better understanding of the failure modes through modeling, laboratory testing and sampling of used castings from various end-users. In addition, we have tested new alloys in a laboratory setting and also in the field to see how they would perform compared to the existing ""standard casting"". Our testing has revealed interesting data concerning the impact toughness, sigma phase content and formation, chloride attack and other performance attributes of various grades of stainless steel when exposed to cyclical temperatures. Interesting results were also obtained concerning the addition of rare earth elements (Niobium, Boron, etc), nitrogen and manganese to a typical ASTM A297 grade of stainless steel. To date, a few hundred test castings have been installed in iron ore pellet plants worldwide. Many of these test castings have already experienced a year or more of operation with good results.INTRODUCTION Cast stainless steels (ASTM A297 and similar) are a main constituent used in the construction of iron ore pelletizing equipment. A typical straight grate pelletizing machine would contain upwards of 160,000lbs of ASTM A297 Grade HH castings. With a typical casting lasting only two to four years, and a casting costing around $35USD, it is easy to see that the castings are a main contributor to the overall cost of producing the pellets. Grade HH and HI (see chemical composition listed below) became the “go to” materials in the 1960’s and very few changes have been made to this day (the exception being plants that switched to flux pellet production in the 1980’s)."

Jan 1, 2015

"Above ground, elevated belt conveyors that span more than 200 m (656 ft) between support points can cross valleys, rivers, forests and buildings with a minimal footprint. The cable bridge conveyor (CBC) is a new type of suspension bridge based belt conveyor system that has long spans between support points. It combines a typical belt conveyor and an improved simple suspension bridge to form a conveyor bridge. The improved simple suspension bridge places a widened bridge deck directly on suspended cables, without the need for high truss towers. The benefits include reduced cost and increase aerodynamic stability compared to conventional suspension bridges. The CBC’s compatibility with current belt conveyors allows versatility in application.A belt conveyor is a ubiquitous material handling system. Conveyors can be placed directly on ground or elevated above ground by support structures. Considerations of cost, design, environment, civil work, safety and process requirements are all factors in choosing the ground-based or elevated conveyors (Zhang, 2014). The ground-based conveyor has the lowest structural cost when the conveyor is placed on an existing road. Maintenance access is easy and straightforward. Compared to the groundbased conveyor, the elevated conveyor can cross existing buildings and structures, give passage to traffic beneath the conveyor, has a smaller footprint and requires less civil work to pass mountainous areas. The drawback of the elevated conveyor is higher structural cost, which can be compared with savings in civil work and other benefits. Conventional elevated conveyors are supported by rigid truss-type structures, with a typical span of 15 to 25 m (50 to 82 ft) between concrete support bents or steel trestles. Triangular truss, using steel pipes instead of hot rolled steel sections, is more efficient. The typical span of a triangular truss-supported conveyor is around 30 to 40 m (100 to 130 ft), while the weight is about 10 percent less than conventional box truss with a shorter span (Jennings, 2013).Long-span elevated structures based on suspension bridge is another step forward in increasing the distance between support bents, reaching 100 to 400 m (330 to 1,312 ft) and beyond. The wide distance between the support points of an elevated conveyor offers unique benefits. The conveyor can cross rivers, valleys, highways, forests and other obstacles with minimal footprint and reduced civil work. The versatility from long-span conveyors makes bulk material transportation over previously impossible terrains now a feasible option."

Jan 1, 2014

By Michael Day, Paul Kos, Scott Brinton

A hydrogeologic characterization and impact analysis was performed for a proposed quarry at the Hitch Rack Ranch located in the Little Turkey Creek watershed southwest of Colorado Springs, Colorado. Continuous multiple channel tubing, nested piezometers, stream flow measurements, and mapping of known seeps were used to define the three-dimensional potentiometric distribution and areas of groundwater recharge and discharge. Water balance considerations, mass balance calculations, and potentiometric response to earth tides and barometric fluctuations confirmed that the granite fracture system in the ridge area of the proposed quarry has an extremely low bulk hydraulic conductivity resulting in minimal groundwater discharge to sustain the baseflow of Little Turkey Creek. The Little Turkey Creek valley forms the major conduit for groundwater flow exiting from the upper part of watershed underlain by granite. This conclusion has significant implications with respect to mine planning to minimize projected impacts to the hydrologic system and preserve this recharge component to downgradient Denver Basin sedimentary deposits.

By R. B. Krog

Airflow patterns in longwall bleeder systems are affected by gob permeability and porosity and the condition of the gateroad entries after mining. As longwall panels are mined, areas of the caved zone and portions of the remaining gateroad entries become inaccessible, making direct ventilation measurement impossible. Mine operators are left to estimate direction and amount of airflow in the gob and the surrounding gateroads. The most readily available locations are where air enters the worked-out area, approved measurement point locations (MPLs) and/or evaluation points such as bleeder evaluation points, intake evaluation points, and the outer bleeder entry. Ventilation models (network or Computational Fluid Dynamics, CFD) can model the gob and the surrounding bleeder system but calibrating these models is difficult without field data. NIOSH installed a tube sampling system ahead of a longwall face and, as the face retreated, released a slug of tracer gas (SF6) to determine transportation times and flow rates in the inaccessible gateroads. The data was used to calibrate both network and CFD ventilation models for a longwall panel. Both model types were used to predict the airflow patterns in the gob and to determine the airflow rate from the longwall tailgate corner back to the bleeders.

Jan 1, 2011

By J. W. Hawkins, K. B. C. Brady, M. W. Smith

The Pennsylvania Department of Environmental Protection (PaDEP) has been issuing surface mining permits since 1984 that authorize remining in areas that will affect preexisting pollutional discharges. These permits require the mine operator to employ best management practices to reduce pollution loading from abandoned mine discharges. A total of 110 completed mining operations with 233 discharges were evaluated to determine the impact to water quality. Overall, acid loading from these discharges was reduced by 7, 200 kg (15, 900 lbs) per day or 61%. Metal loadings were also reduced. For acidity, 49% of discharges improved or were eliminated, 52% showed no significant change and fewer than 1 % got worse. Loading reductions resulted from a combination of reduced concentrations and lower flows.

Jan 1, 2003

By Yoginder P. Chugh

Physical mine modeling and computer simulation techniques were used to conduct technical feasibility and economic analyses of selected pit configurations for surface mining a single coal seam dipping at about 15 percent to a maximum depth of 150 feet. The pit orientations along the dip (dipline), strike (strikeline) and along an apparent dip were studied. Physical models of each pit configuration were prepared first to delineate potential operational problem areas, and also to furnish a data base-for the computer simulation models. A dragline micro- simulation model (Fluor), a shovel/truck simulation model and a cost analysis model were used to conduct the study. Sensitivity analyses were conducted on variables such as the pit width, block length, and angle of dip. The productivity (Cu.yds/hour) of the mining equipment (dragline or stripping shovel) under each set of conditions was used to optimize the pit configuration. Both block length and panel width were found to significantly affect the dragline productivity larger size trucks were preferable to smaller size trucks in the pit configuration along an apparent dip. The cost analysis results showed that the pit configuration along the seam dip was more economical than the pit along the strike or along an apparent dip.

Jan 1, 1984

By Q. Xiao, R. Naidu, W. Li, S. Song

Acid mine drainage (AMD) derives from the oxidation of sulfide minerals, primarily pyrite (FeS2), and is the most severe environmental issue facing the minerals industry. The most common short-term approach to AMD treatment is migration control, such as acid neutralization and metal/metalloid and sulfate removal, through the addition of alkaline materials, including lime (Ca (OH)2), limestone (Ca CO3), gangue minerals and industrial wastes. This requires the continuous input of materials and may result in the production of a vast amount of secondary sludge requiring further treatment and disposal. Addition of chemicals is usually more important in metal/metalloid removal than in sulfate removal unless the sulfate is present in very high concentrations. A more promising long-term strategy for AMD prevention is source control through the complete removal of pyritic minerals and encapsulation of potential risk minerals by coating with impermeable surface layers. This is regarded as the most cost-effective approach, although the mechanisms underpinning this and the implementation procedures are yet to be fully elucidated. It is likely that long- and short-term practices can be combined to optimize the remediation of contaminated mining sites. Some factors such as differing geological and mineralogical characteristics and transportation costs must also be considered for the successful implementation of AMD prevention and remediation strategies. This review also considers some implications for AMD remediation, but the promising bioremediation of AMD is not discussed as it has been extensively reviewed.

By David Kwietnewski

Roadheaders are gaining widespread acceptance in North America as a means for mechanical tunnel excavation. Until recently, almost all of the large-scale North American tunnel projects were excavated one of two ways ? the tried and true drill-and-blast method or with the use of tunnel boring machines (TBM), the hi-tech super-excavators. TBMs enjoyed fast-growing popularity worldwide as TBM manufacturers offered greater technology to meet the demand of the civil tunnel industry in the late 20th century. Recently, though, more tunnel projects, large and small, can thank the brute force of a roadheader for getting the job done.

Jan 1, 2011

By R. Vance, C. Cho, N. Mardi, Z. Qian, K. Prisbrey

One problem with the application of neural networks to the multivariable control of mineral and extractive processes is determining whether and how to use them. The objective of this investigation was to compare neural network control to more conventional strategies and to determine if there are any advantages in using neural network control in terms of set-point tracking, rise time, settling time, disturbance rejection and other criteria. The procedure involved developing neural network controllers using both historical plant data and simulation models. Various control patterns were tried, including both inverse and direct neural network plant models. These were compared to state space controllers that are, by nature, linear. For grinding and leaching circuits, a nonlinear neural network-based model predictive control strategy was superior to a state space-based linear quadratic gaussian controller. The investigation pointed out the importance of incorporating state space into neural networks by making them recurrent, i.e., feeding certain output state variables into input nodes in the neural network. It was concluded that neural network controllers can have better disturbance rejection, set-point tracking, rise time, settling time and lower set-point overshoot, and it was also concluded that neural network controllers can be more reliable and easy to implement in complex, multivariable plants.

Jan 1, 1998