NSW mine tests RFID thesis

Can trucks safely become more productive?

A NSW mine site has begun trialling an algorithm that detects minor variations in underground truck movements that could have a big impact on vehicle productivity over time.

The system forms the basis of a doctoral thesis being developed by CSC Australia's location awareness solutions principal consultant Kobus Du Plessis.

Du Plessis said results of the NSW mine trial would be added to his thesis. He declined to reveal the identity of the mine involved.

"At this point in time it's between us and the mine themselves," he said.

"[The trial is] looking at the potential for improving the turnaround time of trucks moving from top to bottom [in the mine], getting a load and then moving from bottom to top."

Du Plessis' thesis built on radiofrequency identification (RFID) work he had done at other mine sites, notably a collision avoidance system designed for BHP Billiton's Cannington mine in northwest Queensland.

The Cannington system managed how 50-tonne trucks travelled along a spiralling 6-kilometre main tunnel, to a depth of 605 metres.

RFID tags were mounted onto the rooftops of the trucks and scanned at "critical points" by readers.

Data such as the truck type, direction and movement was captured and displayed on ruggedised screens on the tunnel walls.

The Cannington system made underground truck drivers aware of oncoming vehicles - particularly those ascending to the surface with a full load - so they could safely pass each other in the tunnel.

"In coming up with the design, I had to understand how the truck drivers experienced their driving and how they experienced facing having another truck coming towards them [in an area where] they can't pass," Du Plessis said.

"I actually applied for an underground drivers' license to drive the trucks myself to get that full experience."

Du Plessis' current work was aimed at making RFID "not just a tool that could provide information about where a vehicle is" but also one that could lead to productivity improvements in mine sites - without compromising driver safety.

His algorithm used event data generated by RFID tags on mining vehicles to determine the speed of the vehicle and measure differences in time taken to complete a descent-load-ascent roundtrip.

It then calculated a "weighted average" so the mine site could estimate how long trips should take.

Any major differences in the time taken to move between specific sections of the mine could cause an alarm to be raised with a supervisor.

"When there's any deviance from a set variance we will then alarm the supervisor to say, 'Maybe there's something wrong on the road, maybe the driver is tired, maybe there's something wrong with the truck wheels'," Du Plessis said.

"'We don't know what the problem is but we do know that if you would go down and have a look at this particular zone because that's where we detect some changes in speed, you may be able to find there's a rock [on] the road or maybe there's something on the truck you need to check'.

"If, for example, a truck goes down 12 times and back again and it loses two minutes [for each] turnaround time, over a period of months you lose a lot of productivity time of that actual truck," Du Plessis said.

Du Plessis hoped that the trial would uncover how to run the system safely in a mine site and secondly whether it aided productivity of mine vehicles.

However, he also hoped in the long-term that it could be used as part of driverless, automated mining systems.

"That's the ultimate goal," he said.