Many variables must be taken into account when fine-tuning a grinding line, requiring systematic analysis to achieve outstanding results
By Simon Walker, European Editor
One of the most energy-intensive components of mineral recovery, grinding mill operation is subject to a whole range of external influences. Rock characteristics and properties; mineral liberation requirements; grinding media properties; rotational speed; liner design and wear; all of these and more play a role in determining how a mill performs and, more importantly, how that performance can be optimized.
As Omar Arafat and colleagues from Metcom Technologies pointed out in a paper presented at this year’s Canadian Mineral Processors Operators Conference, held in Ottawa in January, “Grinding circuits are complex systems. There are a large number of active variables, and complicated interactions exist between them, which determine the output. The complexity of these circuits warrants a systematic approach to their analysis and improvement of the processing performance.”
Hence the growing requirement for systems that can collect mill operating data and then use these as real-time input for optimizing how the mill functions—not only today and tomorrow, but also over the longer term as operating conditions alter over time.
Building a Mill-circuit Model
As Honeywell pointed out in a recent white paper, “model predictive control (MPC) has become the de facto standard for implementing advanced [process] control (APC) strategies. Ensuring that the implemented base of MPC applications continue to perform well and deliver expected benefits requires a combination of appropriate monitoring technology and best practices for maintaining the advanced control applications.
“However, the initial process improvements and benefits may not be sustained unless appropriate steps are taken to monitor and continuously adapt APC performance to changing conditions,” the company added. “The appropriate steps involve a combination of the right metrics and work processes being put in place to detect, diagnose and remedy any significant performance degradation of the APC application in real time.”
Honeywell’s Profit Optimization suite addresses all aspects of APC and optimization from improving regulatory loop control to optimizing the entire process. For example, its Profit Controller application allows easy implementation of multivariable control and optimization strategies, and provides safe control of complex industrial processes, while Profit Optimizer software integrates with this application to deliver solutions to real-time optimization problems.
As an example, a few years ago the company assisted BHP Billiton at its Cannington silver, lead and zinc mine in Queensland to overcome problems that were being experienced with its autogenous mill. With mill overloading a major cause of plant downtime, the company used its Control Performance Optimizer to create a simulator that would give the mill staff a better understanding of their grinding circuit, as well as developing a mill charge-volume estimator to provide real-time feedback on the state of the mill.
The company described Control Performance Optimizer as an advanced control and simulation software that provides a complete environment for modeling, simulation, design, prototyping and implementation of conventional through to advanced control strategies.
The dynamic simulator predicts 18 outputs including mill weight, power, volume filling, cyclone overflow sizes and flow rates, showed good correspondence with the real system during the validation stage, and can be relied upon to provide a sound base for future control strategy simulation and development, Honeywell stated.
Supplying Particle Size Data
According to Outotec, grinding circuit efficiency is key to improving the overall performance of a concentrator, so good grinding-circuit control is of major importance. This in turn relies on inputs such as dependable real-time particle-size information. The company has a long history of supplying PSI particle-size analyzers to help optimize grinding circuits, and more recently has built on this experience to provide advanced grinding-control solutions.
As an example, its PSI 300 is an on-line particle-size analyzer system for mineral slurries. Automatically taking samples from up to three process streams, it measures P40 to P90 particle size in the range of 25–1,000 µm (500–15 mesh) as well as the sample density and pH. The company added that the equipment can also be used in regrind circuits since it is not sensitive to entrained air in the sample.
Information from particle-size and other online analyzers is used for closed-loop and expert system control that can help reduce variations in process parameters. This leads to better maintenance of the required particle-size distribution while optimizing the mill throughput to provide a steady feed for downstream processes.
Outotec said its advanced grinding control system tailors the most suitable control strategy for each circuit on the basis of process data analysis, available instrumentation and current operating methodologies.
Optimizing in Iron Ore
In a paper presented to last year’s International Mineral Processing Congress in Santiago, Chile, Jianjun Tian of CITIC SMCC Process Technology and colleagues from CITIC Pacific Mining described optimization procedures used on the autogenous grinding (AG) mills at the Sino Iron operations in Western Australia. Dimensioned at 40 × 33 ft, these mills are the largest of their type in the world, with the plant having a design capacity of 24 million metric tons per year (mt/y) of magnetite concentrate.
The grinding circuit consists of single-stage AG milling in closed circuit with cyclones and pebble crushers. The run-of-mine ore is hard, and is likely to get harder with depth, this being a major factor in the decision to use AG mills rather than SAG mills—which might have high media and liner costs. Design capacity for each of the six milling lines is 1,566 mt/h, with a 180 µm P80 product size.
The authors noted that following commissioning of the first milling line, various engineering design, process and control issues were identified. These included the mill power draw being higher than expected, even at part loading, a mismatch between the mill and pebble crusher capacities, difficult mill control because of load fluctuations, and feed line blockages in the cyclones. The mismatch in capacities in particular led to overgrinding, with the mill unable to achieve its design output.
Inadequate pebble discharge capability in the mill was identified as one bottleneck, which modifications to both the mill and pebble crushers addressed. CITIC SMCC then simulated the milling operation using a model that related the feed, mill geometry and operating conditions to the specific energy. This was used for what-if simulations that showed, among other things, that feed and downstream process conditions may mean that the design AG mill grind size may not be optimal all of the time. The authors also noted that, “modeling will be very helpful in providing guidance to the operation.”
Expert Control and Supervision
FLSmidth’s ECS/ProcessExpert control and supervision system provides advanced process control and optimization for minerals plants, including grinding circuits, by analyzing a wide range of signals and actuating automatic adjustments to manage the equipment and process performance. ECS/ProcessExpert measures the mill power consumption, load impacts, mill mass, sump levels, circuit flows, pump power, stream density, hydrocyclone pressure and product quality, using these data to control the feed rate, mill water addition, mill speed, cyclone feed density, pump speeds and circulating loads.
The aim, the company stated, is to improve the mill circuit grinding efficiency, which in turn results in improved energy consumption and throughput. By stabilizing the plant and balancing equipment loads, the system manages and corrects process disruptions and minimizes equipment downtime. In addition, improved circuit availability and utilization help to reduce operating and maintenance costs.
Installed on a standard PC, the system has a graphical interface and advanced alarm and trend capabilities that are easy to use and manage, FLSmidth reported. The system can control any grinding circuit that uses AG, SAG, ball, pebble or rod mills, and is compatible with most other plant-control systems.
ECS/ProcessExpert can also draw on data from sources such as the company’s Impactmeter, which uses acoustic sensors and proprietary software to monitor and interpret load-induced impacts generated inside SAG mills. The system can adjust the mill operating conditions to improve ore reduction and improve energy efficiency, while reducing ball-on-liner impacts and helping to prolong liner life.
In February 2014, CiDRA Minerals Processing launched its CYCLONEtrac classification optimization family of products. The CYCLONEtrac Oversized Monitoring system (OSM) and Particle Size Tracking system (PST) provide reliable, real-time coarse particle monitoring and particle size tracking on individual hydrocylones, the company said, enabling plant operators to reduce process upsets, improve closed-circuit grind and classification, and maximize recovery.
Both are clamp-on systems designed for primary grinding that operate at the individual hydrocyclone level. The OSM system monitors the flow of coarse particles, and enables operators or automated control systems to take immediate action to avoid prolonged periods of lower recovery through the early detection and isolation of pebble events. The PST system tracks a key particle size parameter, providing real-time feedback that can be used to control a closed-circuit grind process.
The company claims that benefits include the ability to identify poorly performing cyclones while providing real-time cyclone rock detection and rock count trending. The systems can also provide real-time operating performance during plant commissioning, as well as help operators avoid prolonged periods of reduced recovery and poor milling efficiency.
Metso’s SmartEar links microphones by the mill (left) to screens and audio output in the concentrator control room (right).
Metso sold its process automation systems business to fellow Finnish company, Valmet, in April, with Valmet now offering these systems to the minerals industry through its DNA range of products. Nonetheless, Metso still offers some valuable tools to help operators monitor and optimize their mills’ performance.
For instance, Metso’s SmartEar is a SAG mill acoustic monitoring system designed mainly to prevent damage and minimize wear to mill liners caused by the impact of grinding balls. SmartEar also identifies anomalous noises and can indicate the presence of large foreign objects in the mill (such as feed chute plates, shovel teeth and so on) or when liners or lifters have come loose. It can also be used to improve grinding circuit control by detecting mill overload and providing an estimate of the filling level.
Consisting of microphones, a signal transducer system and software to process the data, SmartEar provides a filtered audio signal from the mill to the control room, allowing operators to listen to the mill without any plant background noise. It can operate alone as well as interfacing to any control system, Metso told E&MJ.
Meanwhile, the company’s ExperTune PlantTriage software allows experienced control performance specialists to collect and analyze real-time performance data, identify the root cause of underperforming loops, make and prioritize corrective action recommendations, and report progress.
The ABB System 800xA advanced process controller installed at Boliden’s Aitik mine includes the capability to produce Snapshot reports for easy information access.
Mine-wide Data Integration
Earlier this year, ABB released the latest version of its System 800xA distributed control system. This provides customers with a more secure automation environment, the company stated, having been designed not only for new projects but also to support upgrades of older DCS systems running on operating systems like Microsoft Windows XP.
System 800xA is at the heart of ABB’s recently announced Next Level Mining concept, which integrates operations and advanced information technology. As the company pointed out in a white paper, a milling circuit is a complex, multivariable interacting system in which dynamically changing ore conditions and wear parameters pose particular problems for grinding controllers.
By integrating data, ABB explained, the automation system optimizes grinding controls, avoiding the need for constant manual tuning. Data integration is key—the grinding process can not only be optimized within itself, but because it has access to relevant data from the rest of the mining operation, it can obtain far more leverage. Thus, feed-forward data on ore quality and quantity, production rates or market demand can be used to optimize the grinding stage, the company added.
As an example, ABB reported on the System 800xA-based controls that it has provided to Boliden’s Garpenberg mine in Sweden as part of the current capacity expansion project there. The automation platform helps coordinate suboperations including power and automation for the mill drives, hoists, ventilation system and crushers. The company also provided information on the two low-speed dual pinion ring-geared mill drive systems that it commissioned earlier this year at Imperial Metals’ Red Chris gold-copper mine in northern British Columbia. The drives included advanced mill control systems, with features including variable-speed operation, real-time frozen charge protection, a frozen-charge remover function, controlled roll-back, and positioning without a need for an additional inching drive.
Taking the Guesswork Out of Scheduled Shutdowns
MillMapper shows wear patterns on liners and lifters through color coding.
While knowing when to change worn mill liners can be a challenging task, Outotec claims that its MillMapper 3-D laser scanning technology can help to reduce downtime and liner costs, and improve overall grinding efficiency.
In order to optimize running costs, grinding mill operators and maintenance crews need to be sure when to change worn liners. However, Outotec pointed out, there are often a number of unknowns involved, such as not knowing at what stage of the liner life to adjust the mill weight and speed setpoints to maintain grinding efficiency. To help overcome this, the company offers its Scanalyse service, based on MillMapper. The software displays detailed three-dimensional representations of the thickness of all monitored mill liners, then produces a report that assists in forecasting the remaining liner life-span.
Outotec cited an example of the use of this technology at First Quantum Minerals’ Kansanshi copper mine in Zambia. Since opening a decade ago, the mine has increased capacity to the current 340,000 mt/y of copper and more than 120,000 oz per year (oz/y) of gold, with a further expansion project aiming to increase copper production to around 400,000 mt/y.
Kansanshi began using the MillMapper service in 2013, with the initial results indicating that the shell liners on the mill being monitored would last for nearly two months longer than had been anticipated before a change-out was needed. Outotec quoted the lead comminution metallurgist at the mine, Sonny Mwansa, as confirming that the system has made liner management easier while giving the mine increased confidence in creating reliable and accurate operation and maintenance forecasts. In addition, availability has been improved since fewer stops are required for liner monitoring and relining.
|Output from MillMapper scanning can include liner wear trends (left) and wear profiles for individual lifters (right) for comparison against the original.|
“With MillMapper we’ve been able to improve our mill efficiency in terms of sustaining mill throughput over a full mill liner life,” Mwansa said. “In two days, we will be stopping the SAG mill for replacement of the shell liners, but are still milling at a high rate. This is directly attributable to our use of the information provided by MillMapper.”
Outotec stated that by having an independent, quantifiable and reliable system for monitoring high-wear components, management of operation and maintenance performance becomes a lot more transparent. For example, using MillMapper means that Kansanshi can monitor for liner cracks and make an assessment of complete failure risk. Mwansa added that this has been a very useful tool for holding the liner manufacturer to account over liner quality, and that even though the mine gets regular liner reports (with forecasting) from the manufacturer, the mill staff value having an independent, unbiased report from Outotec as well.
Mwansa also noted that things have been much more stable since using MillMapper, with one of the main reasons being improved management of the mill speed and weight over the liner life. “We run with a mill weight control philosophy, where mill weight is derived from bearing pressure,” he explained.
“Previously, we had a bottleneck of not adjusting the mill speed or mill weight that we later realized was causing us to run at higher than our 28% charge filling target. This led to throughput dropping off near the end of the liner life. Using MillMapper information, we now adjust accordingly, and performance is much better.”