A project by Australia's Commonwealth Scientific and Industrial Research Organisation (CSIRO) to automate the transportation of molten aluminum during the smelter process could some day see machines performing tasks without the need for human control.
Researchers at CSIRO's Information and Communication Technologies Centre in Brisbane have been developing components to automate hot metal carriers, the 20-tonne forklift-like vehicles that transport molten aluminum in large containers called crucibles during the smelting process. Although still in the development and trial stage, the hope is to take humans out of many aspects of the production process, improving both efficiency and overall safety. Just as workers have been supplanted by robotics in many assembly lines, automation is likely to penetrate even further into industrial activities
"Eventually, autonomous vehicles will be ubiquitous in such production industries as smelters and mining," Ashley Tews, lead researcher for the autonomous hot metal carrier project, said. "We would hope that our technology will be part of the automated future of these industries."
The project began in 2005 when CSIRO scientists received an end-of-life hot metal carrier from Australia's Comalco Ltd., now part of Rio Tinto Aluminium (RTA), and set about adding both programmable control systems and remote navigation components, including laser range-finders and cameras for navigation, obstacle detection and crucible docking. Since then, researchers have developed higher-level systems to enable the vehicle to conduct autonomously all the operations of a manned vehicle.
Tews said the technology-laden carrier has successfully conducted hundreds of hours of scheduled and unscheduled demonstrations and tests at the Brisbane facility. One of the most significant experiments involved five hours of continuous operation during which the vehicle traveled more than 5 miles and conducted 60 load-handling operations.
More recently, the CSIRO team has begun testing the non-control technology on a manned vehicle operating at RTA's Bell Bay smelter. The RTA test vehicle is not autonomous like the one at the CSIRO site, an important distinction, according to Tews. Instead, it is testing the technology that hopefully will allow a vehicle to eventually navigate an industrial site autonomously.
"We have been developing and implementing the components at RTA since about mid-2007, and the Bell Bay trials of some of the components are continuing" Tews said. "Development is continuing and the challenges of deploying the technology in a smelter are being addressed. The goal is to get the RTA technology to replicate the successful results achieved with the system operating at the CSIRO center."
Tews said the production-environment trials are critical to advancing the program. "Testing technology in a real application environment is invaluable for refining a system," he said. "For example, site layout, site traffic and environmental considerations are naturally different between sites, and the system is being enhanced to generalize across different environments."
Rio Tinto has shown interest in the project. "There have been no official comments, but we have received great support for our trial from our counterparts at RTA," Tews said.
An autonomous hot metal carrier offers many potential advantages for a smelter. "An autonomous vehicle should provide predictable and reliable performance," Tews said. "It should not have days where it produces lower performance than others, nor does it get tired or bored. Predictability, reliability and repeatability allow managers to more accurately predict maintenance and operating expenses. It also reduces unscheduled maintenance and vehicle wear."
Operating costs also could decline after initial capital expenditures. "Although there is no known large-scale deployment of automated vehicles capable of conducting tasks indoors and outdoors, it is envisioned that the initial outlay for automating a fleet a vehicles and a site will be offset by reduced maintenance and operational overhead," he said. "Autonomous vehicles do not require any of the human requisites of scheduled breaks or overtime, and their potential for continuous reliable operation without ramping up and down shift handovers contributes to this."
Vehicles like the autonomous hot metal carrier also can increase productivity and improve the safety profile of an operation. "Autonomous vehicles have the potential to offer more predictable and dependable performance over a human operator," Tews said. "They also have the ability to outperform a human operator, but this is likely to emerge with well-established systems. For example, if there is a heavy workload the vehicle can be set to operate as quickly as is safely possible, which means that it can accelerate and brake more quickly, which leads to greater wear and more fuel use but increased productivity. In less-demanding times, the vehicle can be set to be more efficient and gentle in its operations."
Safety also is a key issue. "Safety is one of the biggest challenges for allowing automated vehicles to operate at a work site with other vehicles and humans," Tews said. "In one sense, a human's intuition and experience allows them to operate safely, even though they cannot entirely perceive the environment around them. On the other hand, safety of operations can be improved with an autonomous vehicle. The vehicle can use onboard and offboard sensors to examine its surroundings in fractions of a second and the software can make instant decisions based on what is detected. By having more information about the surrounding environment, it has the potential to 'see' hazards that might be obscured to a human operator."
One of the greatest challenges for the autonomous carrier has been accurately docking with the crucible to pick it up. "This required the systems involved to be able to position the hook in the hook eye on the crucible within 10 centimeters," Tews said. "We've been able to achieve this with both vision-based and laser-based systems."
That accomplished, the research team is moving forward with additional trials and development. "We are now focusing on creating redundancy in core systems, improved safety and obstacle management, and dependability in the software and hardware components," Tews said. "On the CSIRO metal carrier, we have demonstrated many hours of consistent and reliable operations, thus proving the foundation technologies required for a fully autonomous vehicle."
The CSIRO team sees the technologies developed for its autonomous hot metal carrier having a much wider potential. "CSIRO's automated hot metal carrier comprises a group of technology components, such as the obstacle management system, navigation system, hardware systems, safety systems mission controller and camera-based docking," Tews said. "Many of these components can be independently applied to other vehicles, including autonomous submarines, unmanned air vehicles and mining machines. Our aim is to commercialize much of this technology and customize it depending on the client's application."
The cost and timeline for commercial development depends on the technology component, the application vehicle, the target environment and the client's required ownership rights, according to Tews.
Right now, CSIRO owns the technology on both the in-house autonomous vehicle and the unit on trial at the Bell Bay smelter. "At some point, various components will be commercialized," he said. The technologies could be licensed to other aluminum or industrial companies, but Tews said that depends on many things, including which company or companies pick up the technology and their conditions.
The project has piqued interest in the private sector. "The CSIRO autonomous hot metal carrier is constantly conducting demonstrations to commercial clients and interested parties," Tews said. "There have been several companies interested in various components."
Tews said it is important to note that the CSIRO team is not exclusively in the business of developing a fully autonomous vehicle, but rather the technology and components that allow a vehicle to conduct autonomous operations. This approach could potentially balloon the realm of possibilities in which the CSIRO work could eventually be applied.