How Digitalization has helped solve Complex Problems – A Case Study Digital Soft-Sensor for Prediction of Real-Time Particle Size

Problem Statement

As the Particle Size Analyzer is currently not operational in Agucha Stream-2, the operations team was charging grinding media based on the previous day’s particle size (P80) reports and the current Ball Mill Power. P80 has to be tracked for effective operation of the grinding section.

Solution Implementation

• A soft sensor for P80 was built by modeling the grinding process using the historical process parameter data and daily P80 data from lab reports
• The model was validated by comparing the predicted P80 with the instantaneous P80 from Lab, and was found to have an accuracy of 97%
• The model prototype has been productionized by deploying it into digital control system

Key Benefits

• The model will assist in optimizing consumable usage, such as grinding media and process water addition.
• It will ensure a consistent P80 to the downstream flotation circuit, which will help the operations team in reducing concentrate grade fluctuations.
• The model acts as a soft-sensor thus saving procurement and operating cost of physical sensor.
• It will help in preventing over-grinding or under-grinding, by effective tracking of P80.

This model is an effective solution to bring more insight to the grinding process, help in reducing concentrate grade fluctuations by supplying consistent P80 to the flotation process, and also decrease energy and media consumption in grinding circuit by minimizing recirculation load in Ball Mill.