All Collections
Flotation Optimization
Flotation Optimizer
Flotation Optimizer: How to use it (open-loop)
Flotation Optimizer: How to use it (open-loop)

How to use IntelliSense.io's Flotation Optimizer to generate recommended setpoints

Niel Knoblauch avatar
Written by Niel Knoblauch
Updated this week

Introduction

The Flotation Optimizer is a central feature of the Flotation Application that allows Operators and Metallurgists to optimize their flotation process in real-time. Given the feed rate and material properties of ore fed to Flotation, the Flotation Optimizer employs bast-in-class modelling methods to generate control setpoints that will maximize the metal recovery at the target product grade.

The Flotation Optimizer screen consists of 3 main areas:

  • Feed variables

  • Control variables

  • Performance variables

Generating Flotation Optimizer Recommendations

The following steps can be followed to generate recommended setpoints

1. Update feed variables

The first step is to make sure that the flotation circuit feed variables are updated. This should happen automatically, as most feed values should be taken from the client’s live sensors. The number of feed variables depends on the site and the process's complexity. There’s a drop-down menu for toggling between multiple feeds located on the left side of the Flotation Optimizer (see image).

Take a closer look at the drop-down menu screen below.

Some feed variables might not get updated automatically and might require manual entry. For these, click the “Report new values” button, which will redirect you to a new window where you can add new values from the Manual input widget above by clicking the “Add” button.

2. Review the Strategy

After reporting new values, go back to the main screen to review the current and optimized process performance (see image).
​​

Click the “Show Strategy” button on the right side of the screen, for more details.

You will see a new window, where you can view the Optimizer's Rewards (Performance section) and Control Limits.

These can be edited by an admin user (typically the Metallurgist or Process Engineer, by clicking on the configure button.

This will redirect you to the Optimization Admin screen, where the Optimizer's Value Driver is changed. For more details on how this is done, click the below link. In this article we will return to the Optimizer Screen.

3. Getting recommendations

The Flotation Optimizer runs on a schedule, providing updated recommended setpoints every 30 minutes - and upon changing the Value Driver (see image).

The timer keeps the user informed about the most recent and upcoming updates. This enables more effective planning and a clearer understanding of the flow of information.

The Operators get new recommended setpoints by current feed variables and strategy. Operators can enter these recommended values into the relevant SCADA setpoints fields to optimize their flotation circuit.

4. How to interpret the interface

The interface has been designed so that cells in which the current values differ from the optimized values are highlighted, indicating the need for user action, and simplifying the identification of this condition.

Criteria have been introduced to determine whether the difference between optimized and current values is significant enough to require a change. Cells that require adjustments in this case are highlighted in orange, indicating this need.

These cells represent the ones where the normalised error between the current setpoint and the recommended setpoint is bigger than a configured threshold value.

The number next to the control variable's name indicates how many cells will need to have their parameters changed according to the recommendations. In the previous image, we can see that 9 cells have been identified as requiring a value alteration process on airflow rate, for example, based on the recommendations. In the image below, you can count the 9 cells.
Note that the min & max values per Control Variable are set in the asset metric admin section. See this article for more details on this.

In the optimizer, two situations that can result in data absence are also referenced. The first one is when both the current value and the recommendation are absent.

The second one is when the current values are present, but no recommendations are output by the model, as shown in the following image.

The user can still see not only the optimized values in the performance section but also the corresponding current values and make a comparison between them. In the performance section for each variable, such as "Recovery" or "Concentration Grade," it will display:

  • Current value (what is happening now).

  • Optimizer's RSP value (what will happen if control recommendations are followed).

  • The difference between the two.

For more information, please contact us so we can set up a session to discuss this in more detail.

Did this answer your question?