SAG Grinding Digital Twin


SAG Grinding Digital Twin

The SAG Digital Grinding Twin consists of a digital platform that groups together different tools for prediction, simulation and operational analysis that generates a virtual replica of the mill and is used to diagnose the present condition and assist control strategies in favor of operational stability and maximization of the outcome.




What information does it provide?:

  • Status of complex operational variables to measure such as total charge level (Jc) and ball level (Jb)
  • Main operating KPIs
  • Early alerts of unwanted conditions

System modules

  • Ball trajectory and impact and load profile
  • Liner wear predictor
  • Ball wear predictor
  • Model of total and consumed power for each component of the card (balls, rocks and pulp)
  • Predictor of future power (3 mins) of the SAG mill
  • Level prediction (Jc)
  • Ball evel prediction (Jb)

Platform software components


Analysis of the operating variable of the mill
With this application, the initial business case of the project is generated and allows us to display advanced graphs to show the descriptive statistics of the process


Mill condition estimator
This application estimates the total load level (Jc) and ball level (Jb) and is based on an optimizing algorithm with restrictions and a SAG mill power prediction model


DEM (Discrete Element Method) simulator
This application simulates the physics of the mill and is used to keep the parameters of the mill condition estimator calibrated, based on the current wear condition of the mill linings.


Cloud data integrator
This application allows us to send in real time the process variables of the mill and obtain back the results delivered by the digital twin hosted in the cloud to the local visualization client installed in the mining plant.

digital twin apps

Architecture of the Digital Twin in the cloud:


The Digital Twin is hosted on Google’s cloud, called Google Cloud Platform (GCP). The plant data is integrated using GCP’s IoT Core, where through a secure connection, the process variables of the grinding circuit are sent through these three options:


  • From an App on a local computer in the plant
  • From an App in a private or public Cloud (Azure and AWS), for customers who already have their data in the cloud
  • From an App in an IoT Gateway, for direct integration to the Process Control System (DCS)


Advantages

  • Quick implementation and replication to other mills
  • Greater flexibility and ability to do multiple simulations in parallel
  • Easy and automated update process

Disadvantages

  • Restrictions on the use of some mining companies due to cybersecurity issues
  • Higher latency in results

Architecture in the cloud:


Architecture with local servers (On-Premise)

The "On-Premise" Digital Twin is installed on local servers under the "Black Box" concept, which only has connectivity to the outside through an industrial network port to connect to the local client developed in Labview and has connectivity via 3G or 4G to connect it to the Internet for remote configuration and maintenance.



Advantages

  • Lower latency, which enables integration to the Plant Process Control System
  • Implementation only at the OT level, since it does not make use of the plant’s IT resources

Disadvantages

  • Longer implementation and update time
  • Restrictions of graphic capabilities of the information display and storage interface
  • Restriction on the number of remote clients viewing the results

System Benefits:


  • Increase in mineral treatment of up to 3%
  • Reduction of variability of the SAG mill treatment by 10%
  • Reduces the projection error of the useful life of liners
  • Reduction in Ramp-Up time at the beginning of the lining campaign
  • Reduction of shutdowns or reduction of treatment due to mill overload events