Open Cut » Maintenance & Equipment
The main objective of this project was to demonstrate how subsets of Artificial Intelligence technologies can play a role in reducing the time and cost of processing mine site condition monitoring data, and increasing the number of corrective actions being deployed to the mining fleet to improve reliability and contribute to improving mine profitability.
Seventeen years of lubricant sample data was obtained for a fleet of dozers and processed through the research team's unique data analytics algorithms. The algorithms were designed to provide site maintenance teams with regular updates on fleet health, as well as an ability to derive new and more expansive operational insights from historical data across the life of the fleet. This was achieved through a combination of text and numerical data analytics and visualisations, providing a means to process data from laboratories significantly more quickly than can be completed manually.
The following benefits can be drawn from the application of these new analytics methodologies:
· The short-term / daily analysis process will provide a means for on-site engineers to reduce condition monitoring data analysis workload by:
o identifying samples outside of historical norms requiring further engineering scrutiny,
o moving samples identifying further action yet conforming to historical norms straight to work orders,
o reducing the time required to review and enter work orders into a system;
· A result of this process will be improved effectiveness and quality of:
o on-site analyses,
o laboratory analyses through the implementation of feedback loops;
· The long-term / annual analysis process:
o provides a means to analyse data for a single machine type and components at a single mine site or across mine sites in a single visualisation,
o allows improved benchmarking,
o is significantly quicker and operates more broadly than traditional analytics techniques;
· Economies of scale for condition monitoring analyses;
· A technology and laboratory agnostic approach;
· Text analytics / natural language processing allows high quality insights from professional laboratories to be interrogated in a similar way to numbers.