Mine schedulers have long had to factor in when trucks and shovels need to be taken off the line for maintenance. That has limited the flexibility they have in the mine plan. That plan can also get thrown out if there is an unexpected outage.
BHP vice president of Maintenance and Engineering Centre of Excellence Maria Joyce told Australia's Mining Monthly that the tools at BHP's disposal meant scheduled maintenance intervals were not as important as they once were.
She said those tools made it possible to extend the duration windows between these intervals.
BHP is moving from scheduled maintenance to planned predictive maintenance by using a risk-based approach to monitor and manage critical failure modes of components. Not only does this let it push out maintenance intervals it also helps reduce the risk of unexpected outages.
Add those artificial intelligence and machine learning-driven smarts to the AI-driven fleet management system BHP uses and some real flexibility starts to emerge.
They give the fleet management system more flexibility around when it sends a truck off maintenance. It can instead safely make that decision based on production optimisation.
Joyce said mining was going through an exciting period.
That can seem a little like the Chinese curse: "May you live in interesting times". Exciting can go to terrifying very quickly, particularly given the pace of change the industry is experiencing.
Technology is driving that change.
To keep on top of its fleet maintenance tools, BHP is applying Ironman and APM.
I am Ironman
Ironman, developed by Perth, Western Australia-based company Ox Mountain, applies machine learning to maintenance data of varying quality, sources, format and accuracy to allow a reliability-centred maintenance approach to improving performance.
The system was created for data optimisation.
Using Ironman BHP built a central archive of data and intelligence, which helped it achieve top-quartile truck performance across several of its Australian operations.
Just about every piece of plant on a mine site has sensors recording data of some sort.
Each haul truck has more than 200 sensors looking at a range of things from vehicle speed to various engine metrics.
On its trucks, BHP collects several hundred parameters of sensor data and slices of that data are used by different teams in different ways across the business.
For maintenance, the sensor data collected includes temperatures, pressures, positions of the internal componentry and other control information.
GE Digital's Meridium Asset Performance Management software includes applications including Asset Strategies, Asset Health Monitoring, Reliability & Performance, and Mechanical Integrity. The applications integrate to help optimise asset management.
Through APM BHP can accurately predict when equipment will need maintenance, schedule repairs during planned downtime, and minimise production disruptions and costs.
Included with Ironman and APM is BHP's one SAP system, which helps plan and schedule the maintenance work.
That alone can be a major undertaking.
Consider the challenges of a processing plant shutdown. Not only does the work have to be scheduled the hundreds of workers needed to do that work have to be sourced, and, in many cases flown to site and accommodated. Their tools and the parts needed have to be coordinated as well.
All moving parts (stand still)
BHP has also developed its own tools and in-house capability that uses AI in its decisions along the value chain.
Joyce said the Maintenance Centre of Excellence was also looking at the standardisation of equipment.
"Then there is the challenge of optimising fleet footprints," she said.
"Over time it's dependent on the mine configuration and its life. It's based around safely and reliably making the most out of our assets."
Speaking at a recent conference Joyce said BHP's focus was on productivity and sustainability.
"Productivity remains our single biggest lever to extract more from the resources and assets we have to grow value from the business," she said.
"We see opportunity at the intersection points of maintenance, innovation and technology."
Joyce said BHP had predictive analytics models running across most of its load and haul fleets globally, as well as its materials handling systems.
"These models are developed and maintained by a small footprint of people in our maintenance centre of excellence, which provides real-time, long-range indications of machine health and potential failure or degradation.
"In WA Iron Ore one of the material handling facilities was challenged by ongoing vibration and material handling impacts that threated to shorten the structures' lifespan.
"Through our technical centres we developed a scalable framework where hundreds of gigabytes of sensor data were processed to diagnose and solve for the challenge way faster than any human brain could.
"In fact, it enabled us to identify other locations in the fixed plant structures where we could make changes to prevent the same risk from occurring.
"While a seemingly small example in isolation, this is one of many examples that in aggregate support the WAIO asset to maintain strong reliability and the best cost performance among iron ore producers globally."