COMPUTER PROCESS CONTROL

To conclude this section, a number of operators have reported encouraging results with audiometric measurements on sag mills, yet few, if any, appear to have adopted the technology. Clearly, some further development/evaluation work is required. Gold Recovery

For the purposes of this paper, gold recovery is defined as that combination of unit operations/circuits which concentrates the gold as a cyanide complex into a pregnant solution. This solution is then further treated by an electro-refining or Merrill-Crowe precipitation and pyrometallurgical refining. Under this definition, the most common processing elements are thickening, pre-aeration, leaching carbon adsorption/stripping or filtration.

The control objective of the gold recovery section of a mill is to maximize recovery while minimizing costs, subject to the physical and environmental constraints of the processing line. The notion of on-line measurement and control of recovery requires some form of on-line composition analysis. A particularly important measurement is the gold content in the tailings stream. Today’s technology has not been able to satisfactorily solve this measurement problem.

On the subject of cost minimization, one of the areas of great interest is the measurement of residual cyanide in the leaching circuit. Although this technology appears to be moving rapidly, as discussed in the next section, most of the on-line devices are in a prototype, B — test or early commercial release stage. As a consequence, the majority of operators rely on manual titration measurements.

Taken together, these two limitations indicate why one generally only finds regulatory controls in this area of the plant. (It may also serve to partially explain the operator’s apparent preoccupation with this area.) Turning to the unit operations/circuits themselves, the kinds of regulatory controls one would expect to find in an automated plant are summarized through examples:

* Thickener — underflow density control through manipulation of underflow product flow rate.

* Leaching — ph control through manipulation of the lime addition rate. Direct cyanide flow control.

* Carbon Adsorption — reagents as above. Carbon advance: usually according to schedule and not by measured gold loading.

* Filtration — level control by recirculation. Reagents as above.

* Carbon stripping — materials movement through sequencing controls (i.e., on/off valves). Temperature control through steam flow manipulation. Pressure control through temperature manipulation.

Carbon stripping circuits often require careful and continuous supervision because of elevated temperatures and pressures, and the caustic nature of the solutions employed. It is for this reason that such a control example was chosen for description.

A simplified flowsheet of the carbon-stripping circuit at Hemlo Gold Mines (Larsen et al., 1988) is presented in Figure 7. This is the area of the Hemlo plant where digital control logic is most extensively used. A combination of sequencing, pi regulatory and advanced cascade controls is used. This is best illustrated by briefly discussing the stripping process.

Once the operator has filled the stripping vessel with loaded carbon, he initiates the computer control sequence. The computer starts the fresh strip tank discharge pump and slowly fills the stripping vessel with warm strip solution, from the top, to avoid blinding the vessel discharge screen. Once full, the flow routing valves are adjusted to permit bottom- filling, and the system piping network is allowed to fill with strip solution. This is facilitated by a flow control valve on the discharge line, ahead of the loaded strip tank. When the lines are full, the elutriation flow rate is set for the fresh strip solution. Based on the phase diagram, the back pressure in the system is ramped to a value that will permit a solution temperature of 138 degrees c. As the pressure is increased, a temperature set point is cascaded to the temperature control loop which, in turn, manipulates the steam flow. The temperature set point versus pressure relationship is also based on the phase diagram. Although this process seems relatively straightforward, efforts to run the circuit on manual control during mill start-up were less than successful, and computer control has been used almost since start-up.

When several bed volumes of strip solution have been pumped through the carbon, the computer undertakes to shut down the process using somewhat similar logic. Although not shown on Figure 7, there is some automation on the regeneration kiln involving pressure and sector temperature control. Summary

Process control can offer significant finanical benefit in virtually any gold milling operation. This has been recognized by a number of operators who have moved to exploit the technology, often in the form of the plant- wide, computer-based, distributed control systems. The strategies are based on regulatory and simple supervisory controls. As Rogers (1985) points out, not only do applications of this nature pay for themselves, but they form the foundations upon which optimizing control is predicated. It is of interest to note that many new mills are being designed with plant-wide, computer-based process control in mind (for example, Newmont Gold’s Carlin operations, Echo Bay’s Cove project, Meridian Minerals’ Royal Mountain King project).

Despite this trend, there are still a significant number of the plants being built/operated with only rudimentary controls. The reasons for this are often rooted in fact (there is a shortage of the necessary, technically skilled manpower in the industry) and in fiction (small operators cannot derive benefits from process control).

Although much of the required technology can be borrowed from other countries and from other sectors in the mining community, there is still much to do, particularly with respect to on-line gold hydrometallurgy applications. Related Topics

In the near term, on-line (or near on-line) sensors will have an impact on process control applications. A broader review is provided by MacLeod and Bartlett (1987), but we will restrict our attentio
n to gold hydrometallurgy applications. (A similar kind of review of instrumentation in carbon adsorption processes was recently completed by Mallett, 1986.) In this area, there is little doubt the South Africans have done, and continue to do, the lion’s share of the development evaluation work.

One successful measurement is the determination of residual cyanide concentration. In the present context, the primary application would be in leach circuit control; however, potential exists in tailings detoxification. Two devices appear to have had some degree of early commercial success: CYANOSTAT — Wyllie, 1987; Uys. et al., 1987; Zehnder, 1987; Bull, 1988; and KREGOLD — Ormrod and Fitzgerald, 1987; van Rensburg, 1987. The former is based on an automated colorimetric titration, while the latter is based on ion -sensitive electrode technology. In either case, the more careful monitoring of residual cyanide concentration and the consequent control of cyanide addition would be expected to reduce reagent costs and improve gold dissolution. From the point-of-view of control strategy, the method described by Uys et al., 1987, has some intuitive appeal. They propose that cyanide addition to the leach circuit be adjusted to the mass flow of circuit feed through feed forward control. The changes in ore mineralogy (gold content, cyanicides, etc.,) would be accommodated through the measurement of the residual cyanide level. This signal would be used in a feedback control loop to trim the ratio.

Gold content, the measurement of primary importance, has also received some attention. With respect to on-line gold analysis, two units are apparently capable of measuring low concentrations of gold in solution. These are described by Mallett, 1986, and Robert and Pohlandt-Watson, 1987. Both methods ultimately depend on atomic absorption spectroscopy for the gold determination, but the sample preparation methods are quite different. The principal application of these devices appears to be in the (alarm) monitoring of barren solutions. Turning to near on-line gold measurements, a number of x-ray fluorescence units are available for measuring gold, at relatively high concentrations, in essentially dry solids. These devices are portable and are based on a technology that is well known in mineral processing. The primary application is the measurement of gold loading in carbon adsorption/ stripping circuits. Scitec is marketing a “map Gold Station” that utilizes high-energy radio isotope excitation, improved detection/counting electronics and some innovative software which the company claims will significantly lower the practical detection limit for gold.

For mills employing carbon, a couple of instruments have been developed for on-line measurement of carbon concentration in a pulp. One of the instruments uses the electrical impulses generated when carbon particles strike an electrode to infer concentration (Hulse, et al., 1984). The other is based on the attenuation of ultrasonic signals (Holton, et al.,) not unlike one of the more common on-line particle size analysers. According to Mallett (1986), Mintek has concluded that the former is better suited to qualitative work and that the latter is the quantitative device. From the point-of-view of process control, these analyses would be useful in controlling carbon concentration in adsorption vessels. Clearly, there is an optimal concentration below which gold losses can occur and above which carbon utilization efficiency suffers. In terms of near on-line measurements, Mallett (1986) describes a device that can be used to rapidly assess the activity of carbon from regeneration (or for detecting poisoning effects in the adsorption process). The device is based on an automatically monitored gold-loading test on a dry carbon sample.

The last area to consider is mathematical modelling and dynamic simulation. The utility of these tools lies in a number of areas, all of which can have an impact on process control, i.e.

— Engineer/Process Operator Training (e.g., Bascur and Herbst, 1985);

— Process Analysis (e.g., van Deventer, 1984, Splaine et al., 1982;

— Off-line evaluation of process control strategies (e.g., Gossman et al., 1982, Neale and Flintoff, 1988; and

— as part of the knowledge base in a control strategy (e.g., Exp ert Control, Optimal Control.)

One of the more interesting recent applications of these tools, and one which provides some insight into their application, has been described by Hodouin et al., 1987. This paper documents the development of a model which was used to simulate the carbon-in-pulp circuit at the Vezina mine, in northern Quebec. In a broad sense, the industry should be making a concerted e ffort to employ these tools for many reasons, including the following:

* The current and projected manpower shortage implies high turnover and underscores the need for efficient and economic training programs. The use of simulators for this purpose will ensure minimal plant disruptions and maximum utilization of manpower.

* To equip engineers with the tools to do quantitative process analysis which will, in turn, promote more of the same, inevitably leading to operational benefits.

* Dynamic simulation permits engineers to explore the impact of different operating approaches and control strategies off-line, without disrupting the process. This should be essential preparation for any plant- testing. Conclusions

The rapid growth of the North American gold industry over the past half decade and its probable continued growth in the near future have been, and will continue to be, accompanied by difficulties. Technical manpower shortages and immature technology, to name two difficulties, have stifled process control development. There are plants that make extensive use of process control, albeit at the lower levels, and there are many that rely primarily on manual controls.

The trend is clearly toward increased automation in gold milling, the advantages being more easily recognized by those currently in the design stage of mine development. At this point, the control strategies consist primarily of regulatory and simple supervisory feedback loops, yet the economic benefits of these systems are unquestionable. In comminution, the gold operators can borrow control technology from other gold producing countries or from other sectors of the mining industry. In gold hydrometallurgy, more development and evaluation work in on-line composition analysis are needed. Recent developments in North America, but primarily abroad (South Africa), indicate that such programs are in place. It seems likely that plants designed in the early 1990s will be able to specify process control systems based on this (proven) technology. This will allow the operator to directly address the plant economic optimization problem. B. C. Flintoff, A. J. Neale and J. W. Austin are employed by the Process Technology Division of Brenda Mines. REFERENCES: Anon 1, 1987, Who’s Already Automated, The Northern Miner Magazine, July, page 33. Anon 1, 1988, “A Breakneck Pace,” The Northern Miner Magazine, January, pages 30-54. Anon 2, 1988, “The New Gold Rush in North America,” Engineering & Mining Journal, June, pages 42-46 Anon 3, 1988, “The Future of North American Gold,” The Northern Miner Magazine, pages 55-59. Ara Barsamian J., 1986, Process Control Computer Systems: Spend Money, Make Money, InTech, March, pages 31-38. Argall G. O., 1987, “The New Californa Gold Rush,” Engineering & Mining Journal. Austin J. W., Flintoff B. C., 1987, Production Improvements Through Computer Control at Brenda Mines Ltd., paper presented American Mining Congress meeting, San Francisco, Calif. Bartrum J., et al., 1986, SAG Mill Operations at Kidston Gold Mines, Min. Met. Proc., pages 96-103. Bascur O. A., Herbst J. A., 1985, Dynamic Simulation for Training Personnel in the Control of Grinding/Flotation Systems, Proc. IFAC Symp. Auto. Min. Res. Dev., Australia, pages 315-322. Board R. M., 1982, Instrumentation and Con trol Systmes for Crushing Circuits, in Design, and Insatllation of Crushing and Grinding Circuits, eds. A.
L. Mular and G. Jergensen, AIME, pages 684-698. Brown C. M., Pipke M. A., 1986, Island Copper Grinding Circuit — A Progress Review, Min. Met. Proc., May, pages 122-136. Bull R., 1988, The CYANOSTAT System for On-line Measurement of Free Cyanide Concentration, Paper presented at 20th Annual meeting of Can. Min. Proc., Ottawa, January. Cross H. E., 1973, Metallurgical Considerations in the Automation of Gold Recovery Plants, J. S. Afr. IMM, April, pages 300-306. Flintoff, B. C., et al., 1987, Canadian Trends in the Control of Mineral Processes, paper presented at Copper 87, Chile, November. Flook W. R., Plasket R. P., 1982, Design and Control Practice for Run of Mine Mills in Union Corporation, Proc. 12 CMMI Cong., ed. H. W. Glen, S. Afr., IMM, pages 539-550. Gossman G. I., et al., 1982, The Applicastion of Computer Control to a Gold Milling Circuit at East Driefontein Gold Mine, Proc. 12 CMMI Cong., ed. H. W. Glen, S. Afr. IMM, pages 817-825. Hales L. B., et al., Supervisory Control of a Semi Autogenous Grinding Circuit, Int. J. Min. Proc., V. 22, pages 297-312. Herbst J. A., Bascur O. A., 1984, Mineral Processing Control in the 1980s — Realities and Dreams, in Control 84, ed. J. A. Herbst, AIME, pages 197-215. Hodouin D., et al., 1987, Evaluation of the Performances of a CIP Gold Recovery Process, Proc. 19th Ann. Mtg. Can. Min. Proc., CIM pages 644-676. Holton A. E., 1987, A Prototype Instrument for the Measurement of Carbon Concentration in Pulp, Proc. Colloq. On-Stream Analysis, S. Afr. IMM, 1987, Lecture 4. Hulbert D. G., Barker I. J., 1984, The Control and Optimization of the Operation of Milling Circuits, Proc. Mintek 50, MINTEK, pages 197-204. Hulse N. D., et al., 1984, The Measurement of Carbon Concentration in the Adsorption Vessels of a Carbon-in-Pulp Plant, Proc. Mintek 50, MINTEK, pages 495-500. Jerez J., et al., 1985, Automatic Control of Semi- Autogenous Grinding at Los Bronces, Proc. IFAC Symp. Auto. Min. Res. Dev., Australia, pages 275-282. Larsen C. R., et al., 1988, Process Control in Gold Milling at Noranda’s Hemlo Gold Mines, Proc. 20th Ann. Mtg. Can. Min. Proc., CIM, pages 175-187. Latour P. R., 1976, The Hidden Benefits from Better Process Control, Advances in Instrumentation, V. 31, Part 1, Paper 528. MacLeod J. M., Bartlett H. E., 1987, Chapter 14: Process Control, in the Extractive Metallurgy of Gold, ed. G. G. Stanley, S. Afr. IMM Monograph Series M7, pages 793-829. Mallett R. C., 1986, Design and Operation of Plants for the Recovery of Gold by Activated Carbon: On-line Analysis, Proc. CIP School, MINTEK, pages 18.1-18.23. Neale A. J., Flintoff B. C., 1988, A Dynamic Simulator for Evaluating Control Schemes in an Industrial Scale Rod Mill-Ball Mill Grinding Circuit, Proc. 20th Ann. Mtg. Can. Min. Proc., pages 460-478. Norby S. A., Hales L. B., 1986, Crushing Control at Kennecott’s Ray Mines Division, SME Preprint 86-383. Ormrod G. T. W., Fitzgerald B., 1987, An Improved Leaching Efficiency Detector, Proc. Colloq. On-stream Analysis, S. Afr. IMM, Lecture 7. Parashyniak P., et al., 1988, Computer Based Process Control at the Lupin Concentrator, Paper presented 90th AGM CIM, Edmonton, May. Pauw O. G. et al., 1985, The Control of Pebble Mills at Bufflesfontein Gold Mine by use of Multivariable Peak Seeking Controller, J. S. Afr. IMM, March, p ages 89-96. Robert R. B. D., Pohlandt-Watson, 1987, The Adaption of Analytical Instrumentation for a Mobile Laboratory, Proc. Colloq. On-stream Analysis, S. Afr. IMM Lecture 9. Rogers J. A., 1985, Optimizing Process and Economic Aims, Chem. Eng., December, pages 95-99. Shinskey F. G., 1974, The Values of Process Control, Oil & Gas Journal, February, pages 80-83. Splaine M. et al., 1982, A Model to Predict Changes in the Recovery Efficiency of a Gold Plant, Proc. 12 CMMI Cong. ed. H. W. Glen, S. Afr. IMM, pages 769-778. Uys N. et al., 1987, Automatic Control of Cyanide in Leach Solutions of a Gold Plant, S. Afr. Mec. Eng., April. Van Deventer J. S. J., 1984, Kinetic Models for the Adsorption of Gold Onto Activated Carbon, Proc. Mintek 50, MINTEK, pages 487-494. Van Rensburg D. M., 1987, Practical Experience with the Gold Leach Efficiency Electrode on a Slimes Reclamation Plant, Proc. Colloq. On-stream Analysis, S. Afr. IMM, Lecture 8 Weiss A., 1984, Worldwide Assessment of the Most Profitable Computer Applications in the Mineral Industry, Proc. 18th APCOM, IMM, pages 575-581. Wyllie R. J. M., 1987, Cyanostat: At Last a New On-Line Control for Precious Metals Plants, Engineering & Mining Journal, January, pages 32-33. Zehnder H., 1987, Continuous On-line Monitoring and Control of Sodium Cyanide in Gold Leach Solutions, Proc. Colloq. On-stream Analysis, S. Afr. IMM, Lecture 2. Specs

]]>