Unlock SAG mill capacity with process control optimization

Insight Article by Karl Visnovec, Senior Process Control Specialist based in Toronto, Canada.

 

Intro

The traditional operation of SAG mills relies on the operator to adjust the manipulated variables to maintain the target mill load and power draw. Changes to the ore hardness, feed size or ore specific gravity require an alert operator to adjust the variables while keeping up with all the other demands in the plant. With the operator’s attention frequently drawn away from the SAG mill, the actual mill load would vary around the target, resulting in large corrections to the manipulated variables to bring the SAG back in control. These fluctuations destabilize the downstream flotation or leach circuits, lowering overall plant performance. 
 
The methods on how to manage the changing variables around SAG mills have improved over the years with new instrumentation and Advanced Process Control (APC) systems. Despite these improvements, there is a gap between the available technologies and achieving the maximum potential of the SAG mill performance. The gap traditionally comes from the belief that the base layer controls and existing operation is performing optimally, which is not always the case. Without first identifying missed opportunities for stabilizing the process with existing tools, the addition of new systems can only attempt to compensate for the variability in the SAG mill circuit. 
 
Key strategies for control optimization
 
The following optimization strategies are recommended to achieve high-level process control:
 
  1. Understand the key process drivers affecting the SAG mill performance by assessing the SAG mill performance using comminution models. The objective is to identify the key precursors of the process variables (i.e. ore properties such as competency)
  2. Ensure all instrumentation is calibrated and performing under normal stable operating conditions (i.e. no ore variability in feed composition)
  3. Stress test the controller for sudden changes in ore variability 
  4. Continuously improve, identify and test the controller – the overall objective is to stabilize SAG mill control for high ore variability. A stable SAG mill performance enables operations to push the limit to maximize throughput.
Control strategy implementation

In the above graph, the progress of a well-implemented control strategy is to change from an unstable, highly variable operation to a stable, optimized SAG mill with a responsive controller adapting to dynamic ore and production variability. The optimization steps are:

  • Traditional: The base case control strategy results in unstable operation with highly variable throughput. 
  • Stabilized: Improvements are made over the base case through frequent instrumentation calibration, control tuning and essential updates to the regulatory control. These basic-steps shift the target closer to the maximum SAG mill constraint.
  • Constrained: Further improvements can be achieved on SAG mill control though APC or model predictive control (MPC) control strategy which increases the target setpoints closer to the maximum limit.
  • Optimized: Once the process is working within a tight operational control, realtime insights can be gathered from the instrumentation allowing for further dynamic response. These insights from the measured interactions enable the setpoint target to shift dynamically in response to ore variability, reaching the optimum capacity of the asset while maintaining optimum production objectives (i.e. throughput, lower operational cost). Tools and techniques such as machine learning and Artificial Intelligence (AI) algorithms are enablers of the realtime optimization solution.

 

Instrumentation

The instrumentation installed on a SAG mill is integral to control. For example, the mill load and noise sensors are instruments that provide essential information for an optimized SAG mill operation. Insufficient understanding of the ore characteristics and feed composition can result in an ineffective SAG mill operation. One such example of where instrumentation is believed to be providing the right feedback from the process is the microphone or vibration sensor used for SAG mill speed control. 

The basic strategy works by maximizing the SAG mill power, using a microphone or vibration sensor on the shell to indicate when the impact energy is low, and when the speed can be increased. Alternatively, if high energy impacts are detected, the SAG mill speed is reduced to protect the liners. Although this works well in most scenarios, changes to the feed conditions (i.e. competency, size distribution) can shift the base layer reading for high energy impacts and create false positives or negatives. 

Insights can be gained by analyzing the process variables and identifying signals that indicate these changes in feed conditions. Setting up logic to identify these conditions can allow an operator to adjust the threshold for high energy impacts dynamically. This ensures the SAG mill can continue to operate at higher speeds and increased power draw, without risking damage to the mill liners. Additionally, the instrument only allows the SAG mill to be operated at the limit for the current SAG configuration. To draw maximum power, liner design optimization is a necessary step to target the highest operating speeds after relines, while leveraging the instrumentation to protect the liners (Chandramohan et al., 2018, Chandramohan et al., 2019. See references below).

Advanced Process Control Systems

With APC systems, there is an assumption that any issues in the plant’s control system will be resolved once the system is installed. However, if the in-place base layer control is operating poorly, the APC system is not effective at maintaining the operational targets. Any variability created from the base layer control will have to be filtered, which will cause delays in identifying changes to the process conditions, making the APC less responsive. To achieve the best results with the APC, the base layer controllers need to be reviewed, and any induced instabilities from the controllers, rectified. 

APC systems can also monitor many variables and react quickly.  Because of this, they are frequently used to correct for upset conditions. However, the correct approach should be to first look at options to limit the behaviour of the equipment or scenario that is causing the upset. By using the base layer controls to stabilize the process, the need for compensations by the APC system is reduced or eliminated.

An example of a missed opportunity is when ancillary equipment are run in an ON/OFF mode, instead of using proportional control. See case study here.

Conclusion

Advancements in instrumentation and APC systems can unlock significant capacity in system and SAG mills by providing additional feedback and rapid responses to the process dynamics. While assessing the implementation of new technologies, a performance audit of existing controllers and instrumentation can help identify new methods of improving the existing base layer controls. By taking the time to review the current state of the control systems, operations can ensure they are building the new technology on a sound foundation. Failure to stabilize the process using base layer controllers can result in lost potential (i.e. throughput, lower operating cost). If the process cannot be stabilized, the SAG mill cannot be optimized.

To achieve the maximum potential of the SAG mill (installed power and resulting throughput), base layer controls in the existing operations need to be configured to minimize the impact of process variability on SAG mill stability. This is possible when the instrumentation is providing reliable feedback from the process, the controllers are tuned to respond to the process variability before new systems – such as APC – can realize their full potential. 

SAG mill optimization work can be carried out remotely, with the implementation of remote connectivity, open collaboration with on-site personnel, and access to real-time plant data - this allows for lower cost implementations, faster execution rates, and increased throughput from the SAG mill. 

For more information, contact Karl Visnovec.

 

References:

Chandramohan, R., Braun, R., Lane, G., Hollis, K., Johnson, Le, J., and D. Baas,  (2018). SAG mill optimization and increasing throughput at the Phu Kham Copper-Gold Operation, 14th Mill Operators Conference, AUSIMM, Brisbane, Australia (View paper here)

Chandramohan, R., Lane, G., Pyle, M. and R.Whittering. (2019). SAG Mill Liner Selection to Maximize Productivity, Procemin-Geomet 2019, Santiago, Chile. (View paper here)

;