Process Control in metallurgical plants—From an Xstrata¹ perspective

Abstract

Some history and modern practice of Process Control in metallurgical operations is reviewed. Clearly the early deliverables from the pioneer days in the 1950s through to the 1970s and early 1980s were under-appreciated. The discipline has since grown into a more visible, sophisticated and accepted practice as a result of the assembly of appropriately recruited and trained individuals and teams, who have successfully negotiated deliverable projects that impact all metallurgical performances beyond early milling processes. The skill set in these individuals and teams essentially includes organisational behaviour in addition to their specialist technical attributes. A strong network to internal and external specialists and experts is essential. Furthermore, instrumentation and control technology has improved immensely. The challenge in the current modern practice is to win support of senior management in operations for the project cost, schedule and deliverables of Process Control. Once gained, this acceptance then amounts to the logistics of project scope and delivery—a track record well-demonstrated by the Xstrata Process Control Group.

Introduction

The objective of the third plenary lecture is to present the state of the art of applied automatic control and information processing in one or several MMM fields (physical metallurgy, mineral processing, extractive metallurgy, and mining). Mineral processing and extractive metallurgy are the selected MMM fields for this discussion.

In this paper, automation is understood in a broad sense including: measurement and instrumentation; process modelling and simulation; process monitoring and data reconciliation; data mining and multivariate statistics; fault diagnosis and fault tolerant control; Process Control; monitoring of product quality and control performance; off-line and on-line process optimisation; maintenance scheduling and production planning; as well as AI methods like: expert systems, neural networks, fuzzy control.

To address as many of these areas of automation as possible, this paper divides into the following sections:

• Introduction and Process Control definition, objective and cycle

• Elements necessary for successful Process Control in mineral and metallurgical plants, covering:

  • People—control/process knowledge

  • Tools—instruments; systems; technology

  • Actions—support, management, technology transfer

• XPS Process Control philosophy summary

• The ‘near term’ outlook

• Conclusions

Examples and case studies are briefly presented throughout the paper to illustrate the points being made. Acknowledgements and references follow after the conclusions section.

McKee (1999) states:

“Process control is a broad term which often means different things to different people. For the purposes of this (AMIRA) review, process control is considered as the technology required to obtain information in real time on process behaviour and then use that information to manipulate process variables with the objective of improving the metallurgical performance of the plant. This was, in fact, the objective of the early (metallurgical) control systems.”

Furthermore, he goes on to state:

• “Process Control is a long-established technology.”

• “Within the minerals industry, single loop pneumatic controllers became commonplace in the 1950s.”

• “By the mid 1960's, process control was seen by the mineral industry as a technology which offered a great deal in terms of potential improved plant performance.”

• “Work on flotation control systems began in the early 1970's based upon the availability of on-stream analysis data.”

• “Although the flotation control task was inherently more difficult than for grinding, there were many promising flotation control systems around the world in the mid to late 70's” (and early 80's).

“Thus, all the elements necessary for successful process control in mineral plants were known and largely available (over) 20 years ago, and were:

• Instrumentation

• Control hardware and software

• A growing understanding of process behaviour and dynamics

• Operator interfaces providing communication between operator, the system and the process

• An understanding for the absolute necessity for good instrument maintenance and the capability to perform it

• An understanding of the need for skilled personnel to develop/maintain control systems

• Proven successes of many systems

• Enthusiastic support of management for process control

• The active and productive involvement of research groups working with enthusiastic operations personnel.”

Although it is now several years later, these key elements have been selected to be discussed in this paper, as they are also seen to be very relevant to the approach of the XPS Process Control Group.

The goal of the XPS Process Control Group is plant Process Control for Operational Performance Excellence. The key elements to achieve this are summarised and addressed in Section 2, under the following sub sections:

• People—control/process knowledge

• Tools—instruments, systems, technology

• Actions—support, management, technology transfer.

More recently, Thwaites and Løkling (2002) stated: “Installed instrumentation calibrated properly with appropriate resolution deadbands, and first order filters, is the basis for good proportional, integral and derivative (PID) control. This offers great potential to plants in reducing variability, opening up opportunities for optimisation through increased throughputs or reduced consumables (e.g. power, steel, reagents, etc.). Tools like Expert systems, Model Predictive Control (MPC) Multivariable controls (MVC) with LP (linear program) optimisers, and even “higher-level” (cascade) PID controllers working on analyser information, provide the means to accomplish significant returns, but these are all dependent on a good/reliable instrumentation base and robust regulatory control.” Fig. 1 is a general diagram, used for many years, to illustrate the overall Process Control objective for any key variable. Measure and understand the initial variability, stabilise and then optimise to the constraint of the process. Typically throughput benefits will come from an increase in the setpoint, and consumable savings will come from a decrease in the setpoint. Once these have been achieved it is important to ‘maintain the gain’ by ensuring the changes are robust, thus benefiting the plant.

In the process of stabilisation, it is very important to consider ‘the whole loop.’ ABB (Forsman, 1998) illustrated this very well, as shown in Fig. 2. Are there any loops in manual? There can be several reasons why controllers are in manual, for example: instrumentation; tuning; operator training; saturation and ability of control actions to influence the control variable, etc. Utilising tools like ExperTune (in 1999, selected by Control Engineering as one of the best products of the year), at the commissioning stage, helps to identify the key issues preventing robust loop control.

Fig. 3 illustrates the critical mill bearing pressure measurements which required a basic filter so that the data is usable for control and better represents the real process issue. For sometime, following plant commissioning, all process signals were unnecessarily ‘noisy’ and it was not surprising that much of the plant's control was manual! Alarming in this plant was also ineffective and poorly commissioned.

Additionally, Process Metallurgists/Engineers also need to be aware of ‘data compaction’ in the production management information system (PMIS). Over compaction, as over filtering, minimises the ability to ‘see’ the real process dynamic response. The XPS Process Control Group is often requesting less compaction, while the IT support is often interested in maximum (and default) compaction. Similarly, PMIS systems should also, but often do not, capture important setpoints and outputs.

Referring back to Fig. 1 then, Process Control objectives can therefore be summarised as the management of tools and resources to: minimise process variation; optimise process performance; minimise (variable) costs; control product quality; and maximise safety.

In general a well established hierarchy (Jones, 1994/1996), as shown in Fig. 4, is followed. Economic returns considerably increase beyond loop regulatory control.

While process and area optimisation can have considerable financial returns, it is well recognised (and reported), by those practising Process Control, that this can only be achieved by having a robust and solid regulatory lower level.

McKee (1999) states:

“Svedala (metso minerals) advocate a systematic approach to process control:

• start by defining what is important through an analysis of the process (an audit) to determine what is required and the level of control;

• understand the process and ensure the basics for good control are in place (e.g. instrumentation);

• match the installed system to the capability of the people available;

• recognize the on-going commitment required for success;

• with regard to the actual control, consider three levels as follows:

  • basic PID—via DCS or equivalent;

  • supervisory (cascade control, dead-time compensation, etc.—via DCS or equivalent);

  • optimising (e.g. expert system, adaptive)—via a package such as G2 or implemented within the DCS.”

However, it is important to understand that Process Control will not correct inherent design, instrumentation related flowsheet, and actual flowsheet problems in a plant.

‘There is a need to determine, and if necessary correct, the condition of the plant as a pre-requisite to control development. A good example is the importance of classifier operation and its effect on comminution circuit performance. Techniques exist (plant sampling, modelling and simulation) to audit the actual plant operation. Correcting plant limitations should be seen as a first step in the control approach.’ (McKee, 1999)

The hybrid Process Mineralogy group at XPS was started up in 1997. The key deliverable from this young discipline is to reliably formulate and demonstrate the optimum flowsheet for the processing of a given ore body (Lotter et al., 2002, Lotter et al., 2003). This hybrid approach of sampling and statistics, geology, mineralogy and mineral processing (Fig. 5) outperforms conventional mineral processing because it describes the flowsheet in terms of representative sampling and minerals instead of assays. Ongoing improvement in the concentrator operations, for example by Process Control projects that reduce variance, is thereafter benchmarked with Statistical Benchmark Surveying. The Process Control is easier to implement when a concentrator operation is on a Process Mineralogy platform, because the flowsheet allows for more responsive controls.

“A good implementation requires a well defined operating philosophy and an associated control strategy.” (McKee, 1999)

The XPS Process Control Group continually make the point that Process Control involvement during the early engineering stages (of a project) ensures effective start-up and the fastest turn-around to on-line process optimisation. Fig. 6 shows the Group's ‘effective Process Control cycle’. This also fits in well with the Company's (Xstrata Copper Canada) Six Sigma stage gate process (also shown in Fig. 6).

The Kidd Metallurgical Site Montcalm project was taken through the full Six Sigma stage gate procedure, ensuring all specified design criteria were met before proceeding to the next stage. This required the Process Control plan (Gillis & Lacombe, 2002) to be developed very early on in the process. Piping and Instrument Drawings (P&IDs) were first developed during 2002 and a detailed Process Control philosophy was written up as well. This philosophy detailed every loop, the expected flows, setpoints, operating ranges, limits, alarms, interlocks, etc.

The control philosophy generally specified only lower level (regulatory) loops with some cascade loops included where necessary. The approach taken was to have the regulatory loops ready for use at circuit start-up and to develop higher-level controls later as required. Following commissioning all control loops had been inspected, optimised and documented with the ExperTune software package, thus allowing the Process Engineers and Operators to focus on optimum process beneficiation, i.e. Operational Performance Excellence.

The XPS Process Control Group are often called into Plant processes after start-up to ‘correct and fix’ poorly commissioned process controls, sometimes after a considerable time period following the new process commissioning.

Fig. 7 shows a recently modified SAG circuit (after several millions of dollars of capital investment) showing a tremendous variability of the feed tonnage (lower trend; S.D. of 9.7 tph on an average of 131 tph) and bearing pressure—indicating ‘mill load’ (upper trend; S.D. of 24 kPa on an average of 4454 kPa) following ‘commissioning’ of the new equipment and few basic controls. Needless to say, this ‘fundamental and critical control’ affecting the whole mill, was soon switched to ‘manual’ placing onus on the operator to maintain mill throughput while not exceeding any of the equipment constraints! Much of the operators’ time was then consumed with this task.

“In grinding, control of AG/SAG mill circuits is the dominant area. While some systems have emerged which provide a reasonable level of control, there is still much not understood about the dynamic behaviour of these mills, and there is considerable scope for further development.” (McKee, 1999)³