Grinding Case Study

Home 5 Case Studies 5 Grinding Case Study

Client: Pustynnoye Plant of JSC “AK Altynalmas” (formerly Aktogay)

Project Objective

JSC AK Altynalmas is a fully integrated mining company operating in the Republic of Kazakhstan. Aktogay (Pustynnoye) is one of the Company’s sites where solutions were much needed for debottlenecking the grinding circuit. The IntelliSense.io team was responsible for the following key tasks: reducing mill shutdowns through real-time predictions of ball charge and liner states, reducing the frequency of SAG mill overloads, and increasing the controllability of the grinding circuit.

Challenges

  1. Limited knowledge of changing ore properties lead to;

    1. Frequent mill overloads

    2. Reduced throughput

  2. Instrumentation on the mills is not sufficient to run the mill with optimum ball charge and knowledge of the liner state, leading to;

    1. Frequent mill stops for inspections

    2. Non-optimal running conditions.

  3. Traditional sensors cannot be used within the mill due to the nature of the process.

Solution

The IntelliSense.io Grinding Circuit Optimization application has been piloted at the Aktogay Plant of JSC “AK Altynalmas”. A Digital Process Model of the SAG mill was created using first principle modeling and enhanced with machine learning algorithms – the combination which enables improved monitoring, performance prediction, and optimization to increase profitability. There are four key models involved in the Grinding application:

The Ball Charge model virtually monitors and predicts ball charge to keep it in the optimal range for throughput and grinding effectiveness without having to stop the mill for costly and hazardous inspections.

The Liner Wear Model helps in maximization of the liner operational life by providing information about current liner wear and outstanding capacity in hours and tons of throughput. This allows the planning of maintenance and replacement activities with more precision.

Fig 1: Screenshot of the Flow (Cyclone Underflow) to the Ball Mills
created through the Material Model using machine learning.

Fig 2: The Grinding app being used on site in Kazakhstan.

Fig 3: The graph above shows correctly predicted overloads in advance of the overload

Fig 3: The Operator control room on-site at Pustynnoye (formerly Aktogay)

The Material Transport Model was configured to provide visibility on streams that cannot directly be measured – thereby providing valuable insights about the circuit to operators. Outputs of the Material Transport Model include flow rates, as well as the solid/liquid ratio in slurry streams. For example, the total solids flow rate to the Ball Mills (which cannot be measured directly) is one of the key deliverables of the Material Transport Model. Material residence time is another example of valuable data which was previously unavailable to the operators.

Additional outputs include the mill residence time, the sump discharge flow rate and the total process water added at various points along the comminution circuit. The Material Transport Model also compensates for sensor failures.

The Overload Prediction Model is a real-time AI model that predicts the probability of a mill overload taking place over the next 10-20 minutes, allowing gentle early interventions to prevent SAG mill stoppages. This translates into greater and more consistent throughput.

This model is trained on historical Mill feed and performance data, so that it knows how the Mill responds to changes in solids feed rate, dilution water, material properties and control variables like mill speed and ball additions. It then uses this digitalized knowledge to predict how the mill will respond to current conditions – and you are alerted of a potential mill overload long before it happens.

Benefits

The following key benefits of the Grinding Application are realised and validated in collaboration with the production department of the Aktogay Plant of JSC “AK Altynalmas”:

  • Reduced mill downtime via elimination of stoppages for liner and ball charge inspections
  • Increased throughput via reduction in overload frequency
  • Increased controllability of the grinding circuit via increased visibility of streams

     

”by maintaining productivity and reducing shutdowns the economic benefit achieved a 1% throughput improvement or $1.3M””

Zhanar Amanzholova

VP of IT & Corporate Development, Altynalmas

Mine to Market: Value Chain Optimization

Powered and connected together by the brains.app platform, the Grinding Optimization Application is one of a suite of real-time decision-making applications that uses Artificial Intelligence (AI) to optimize each process; from mine-to-market.

Our Material flow model connects these applications together to drive even greater efficiency gains.

The IntelliSense.io Application Portfolio