The Problem
A large iron ore operation in Western Australia aimed to sustain long-term benefits through pit-to-port optimisation. However, they faced challenges related to material size affecting their dig-rate, crusher and processing plant throughout. As is commonly the case with optimised capital investment strategies, the operational bottleneck changed across the mine value chain over the years. Uplifts were achieved by being able to switch optimisation target/focus as the constraint changed over the years.
They also sought a deeper understanding of the likelihood of “slower to crush ore” entering the crusher and wanted to plan around geological domains and blend accordingly with ore sourced from other blocks.
Additionally, when the bottleneck switched to digrate, the drill and blast optimiser target was switched from “minimise crusher downtime target to maximise instantaneous dig-rate whilst simultaneously maintaining constraints on costs and plant throughput.”
The Solution
After careful evaluation of several pit to port solutions the company selected PETRA’s MAXTADrill&Blast, an advanced solution designed to optimise blast design and improve downstream performance.
This digital, cloud-based, and sustainable pit-to-port optimisation solution provided engineers with the capability to balance downstream productivity enhancements against any associated cost impact. They could analyse the effects of many factors such as reducing burden and spacing, changing explosive density, altering product type, increasing blast size, and changing blast direction which would ultimately affect digrate, throughput and yield.
By leveraging data science, engineers could use historical performance data to inform decision-making, maximising the learning potential for fresh staff members and reducing the need for existing engineers having to spend long hours to create optimised blast designs.
Implementation Details
To effectively implement the solution, extensive sampling was conducted, amounting to tens of millions of tonnes. Data fusion ore tracking was used to create digital “twins” that correlated crusher downtime with geology and drill and blast designs.
The downtime prediction model involved automated parameter tuning and mathematical ore tracking. Hundreds of millions of tonnes of ore data, including blast design and categorical block model variables, were tracked to the crusher.
Two-phase geological domaining, combining supervised and unsupervised machine learning (ML) techniques, was employed. Clustering algorithms like k-means, spectral, and agglomerative clustering were optimised using AutoML.
The oversize downtime prediction models included a classifier neural network that predicted the probability of downtime as a percentage likelihood and regression models utilising gradient boosted decision trees to estimate downtime duration.
The selection of MAXTA Drill&Blast models was optimised for each specific implementation. Data and ore tracking parameters were automatically corrected and calibrated using machine learning techniques.
Engineers were given access to a cloud-based portal with a purpose-built application for drill & blast simulation and optimisation whereby certain blast levers would supply various levels of confidence allowing them to generate effective blast designs in a simple and easy handbook-type interface.
The application also featured a model comparison chart where different machine learning models such as Throughput, Digrate and Cost could be pitted against each other to help find the best blast design patterns.
Results and Benefits
The implementation of MAXTA Drill & Blast produced several notable results and benefits for the iron ore operation:
- Successful prediction of oversize materials, aiding proactive planning and optimisation efforts.
- Average overall improvement of over 5.5% in downstream processes whilst reducing drill and blast cost.
- Comprehensive insights across multiple areas through simulations conducted on more than 15 orebodies.
- Three-year period simulation conducted from 1st January 2017 to 1st January 2020, enabling thorough analysis.
- Estimated annual financial benefit of over $450 million, proving a significant return on investment.
Through the implementation of MAXTA Drill & Blast, the Western Australian iron ore operation achieved significant improvements in efficiency and productivity. By optimising blast design and using data science, the operation successfully predicted oversize materials, reduced costs per tonne, and gained millions of added tonnes per year.
The comprehensive simulation capabilities allowed the engineers to analyse and evaluate various design options, empowering both specialists and non-specialists to interact with the digital twin simulations. The successful outcomes of this implementation highlight the value of advanced optimization solutions in the mining industry, enabling significant improvements in productivity and profitability.
