Efficiency gains for Rio Tinto's Pilbara rail operations

Over the last few years mining heavyweight Rio Tinto has worked hard to make significant efficiency improvements to its Pilbara rail network. A key contributor to this success was the optimisation of train loading operations at Yandicoogina.

An essential part of their integrated iron ore production process, Rio Tinto’s rail system directly affects the company’s capacity to blend ore to customer specifications. Linking twelve mines with three shipping terminals, this infrastructure is the largest privately owned heavy freight rail network in Australia. Spanning some 1400 km of track, it is responsible for moving 220 million tonnes of ore from Rio’s Pilbara mine sites, through some of the harshest and remote country on the planet, to their port facilities in Dampier and Cape Lambert.

Each train is operated by a single driver and comprises up to 234 ore cars. A fully loaded train weighs in at around 29,500 tonnes and is about 2.4 km in length. On average, the network’s cycle time is 28 hours, with a train movement every 25 minutes along the line. Due to the weight and length of each train, three locomotives are used to pull the load at the front and two to push at the end of the train. With such an arrangement, the correct loading of the cars is critical. While consistently loading each car with as much iron ore as possible yields optimum network efficiency, it is important not to exceed the car’s maximum load capacity. Overloading places stress on critical components, causing failures that can lead to train derailments. On the other hand, under-loaded cars represent wasted capacity and poor network efficiency. Surprisingly, lightly loaded cars that are sandwiched between fully loaded cars can be lifted off the rails, also leading to derailments.

Figure 1: Rio Tinto's Pilbara rail network.

Train derailments have a catastrophic effect on mining operations because they damage the tracks and, depending on where they occur, can halt rail traffic from multiple mine sites to port. Repairing the line after such an incident and getting the train back on track can take days. If, in the meantime, sites can’t load and dispatch their final product, their stock piles reach maximum capacity, forcing them to shut down their processing plants and mine production. So, apart from the obvious safety issues and direct costs associated with derailments, the financial impact in lost revenue is counted in the millions.

Mine-to-port logistics operations are of such importance to Rio Tinto that, in 2008, they launched a major project called Drumbeat to optimise the efficiency of their entire supply chain. As part of this strategic project, mine sites were encouraged to improve the speed and accuracy of their train loading operations.

Responding to the challenge

The team at Yandicoogina, recognising Drumbeat as an improvement opportunity, responded to the challenge. They decided to target a longstanding issue responsible for heated debate between the site and port operations. The issue related to a discrepancy between car load weights measured and estimated during train loading operations at the mine site and those measured at the port. The mine operates two train loaders; one had a weighing system with a 2% accuracy, which introduced a variability of around 2.3 tonnes per car. The other relied on a weightometer that provided the weight of ore loaded on each train. This data was then divided by the number of cars to obtain an average weight per car. The port operations, on the other hand, have a facility to weigh each car individually, giving a better indication of the ore weight per car. Unfortunately, by the time the train reached the port, was weighed, and the data fed back to the mine, 24 hours had passed, during which another five or six trains had potentially been loaded incorrectly. The net result was that, while the site was targeting 116 tonnes of ore per car, they were actually loading anywhere between 110 and 120 tonnes per car and were seldom achieving their target weight. The team realised that the only way to improve their loading accuracy was by weighing the wagons accurately at the point of loading, rather than at the port.

A project team was assembled to tackle the problem and to determine the best way forward. Comprising a number of key staff, including process engineers and the site’s production superintendent, the team identified a number of ways of improving Yandicoogina’s train loading performance. The key to delivering these improvements was their ability to accurately measure the weight of each car before and after loading, and to provide the operator with rapid feedback as to how well he was loading the cars. Because loading speed was also an important consideration to operational efficiency, weighing had to be accomplished dynamically while the cars were moving through the loader.

Working with an external engineering firm, the project team identified two of their existing suppliers that were capable of implementing such a system. Proposals from both companies were evaluated. The key selection criteria for the system were accuracy and ease of integration into the existing train loader’s automation system. The team gained management approval for the project’s capital expenditure based on an average gain of one extra tonne of ore per car.

Choosing a supplier and implementing the system

The chosen weighing system for each train load out (TLO) is based on two separate Schenck Process Multirail train scales. Sang Nguyen, the site’s Planning & Engineering Superintendent commented, “The fact that Multirail was already being used successfully in many European ‘legal-for-trade’ applications with an accuracy of 0.5% made the system appealing to us.”

Figure 2: The Multirail train weighing system.

The ‘tare’ scale located on the inbound side of the TLO provides the weight of each empty car heading in for loading. On the outbound side of the TLO is the ‘gross’ scale that measures the weight of each loaded car as it leaves the load out. The Multirail electronic modules are connected to PCs that are responsible for logging the car weights. Each car is identified using an RFID tag and then its empty and loaded weights are logged. The scale PCs are linked to the TLO’s Modicon Quantum PLC via a Modbus/TCP ethernet network. The PLC interprets the weight data and reports back to the operator in real time via the TLO’s Citect SCADA system. This allows the operator to make instant adjustments to the way that he is loading, and the system provides site production staff with valuable statistical data as to loading performance. Because the track scale PCs are linked to the internet via the company’s corporate network, Schenck Process technical staff are able to access the system remotely, providing Rio Tinto with commissioning and troubleshooting support, along with remote software updates.

Nguyen explained that, “This installation is seen as an important R&D project within Rio Tinto. As it’s the first of its kind in Australia, it was a bit of a risk on our side but has definitely paid off. There were several innovations that we came up with during the installation and commissioning of the Multirail system that will be carried forward into future installations,” said Nguyen.

System usability and benefits

Gerry West, Yandicoogina’s Fixed Plant Production Superintendent is responsible for the day-to-day running and efficiency of the mine site’s equipment and processes. Gerry first arrived on site just as the new train scales were being installed and commissioned. Commenting on the system from a user’s perspective, Gerry said, “The Multirail track scale implementation was a very successful part of the Drumbeat project because it delivered on its promises. It has enabled us to control our loading within much tighter limits, allowing us to increase the average load per car by two to three tonnes.”

Gerry and his team use the new system to collect valuable statistical information that allows them to analyse the site’s loading performance and compare it to data measured by the port operations. He said that before the system was installed, data provided from the port indicated that 13% of their cars were being overloaded. Since the Multirail scales have been installed that number has dropped down to less than 5%. Statistics fed back from the port also show that the standard deviation of tonnes per car has improved dramatically.

Project outcomes

Rio Tinto’s management were extremely pleased when they discovered that instead of the projected one extra tonne per car, Yandicoogina were able to achieve between two and three extra tonnes per car. This shortened the system payback period to just a few months.

The discrepancies between the estimated tonnes per car loaded at the site and those reported from the port were a major concern for Mark Rodgers, Rio Tinto’s General Manager - Yandicoogina. The last thing that he wanted was for his site to be responsible for a derailment on a critical part of the line that could disrupt production at multiple mine sites. This type of event is considered within the group as ‘major’ and is said by staff to be on par with the impact of a tropical cyclone. So, not only did the project improve the company’s bottom line through the improved efficiency of Rio’s Pilbara rail logistics, it also provided significant peace of mind to the management team and even helped the company reduce its environmental footprint.

Figure 3: Train scales in operation.

Commenting on the impact of the project Mark said, “From an overall perspective, we are now probably the most consistent operation with respect to tonnes per car.”

“We have a number of different weighing systems throughout the business and it gives us a lot of confidence to know that ours is on track. I used to constantly be questioned about having the right number of tonnes per car.” He said. “We’re always trying to squeeze out as many tonnes as we can so a key benefit of the system is that it has helped us to achieve that.”

By Peter Newfield, Marketing & Technical Support Manager, Schenk Process