June/July 2013

The power to plan

New tool for evaluating cave economics helps Rio Tinto at the prefeasibility stage

By Correy Baldwin

Rio Tinto's Resolution panel caving project is set to become one of the world's largest underground copper producers | Courtesy of Rio Tinto

Rio Tinto expects that 25 per cent of global copper production will come from underground mines by 2020. “It’s a step change for the industry, from the traditional open pit to a move underground,” says Allan Moss, general manager of Rio Tinto’s Underground Technology Centre. “We’re looking at upwards of 100,000 tonnes per day from underground. It forces us to go for economies of scale, and the only mining method that will give us the economies of scale we need is caving.”

Moss says cave planning requires an unprecedented amount of data processing. “We not only need information on the ore body but on the surrounding country rock,” he explains. “For example, in one of our projects where the ore body is 2,000 metres below surface, we need to be able to characterize eight cubic kilometres of rock in order to properly assess cave behaviour, and ultimately determine economics.”

With an almost infinite number of possible mining sequences in such large projects, the decision whether or not to develop an ore body has become incredibly complex. To help speed up the assessment, in 2006 Rio Tinto approached professor Hylke Glass, the Rio Tinto Professor at the Camborne School of Mines, about developing a smart sequencing strategy. The resulting study led to the creation of software called CavePlanner.

“The intent of the software is to get a feel of what the footprint of the cave would look like and what sort of sequence we could adopt,” Moss says, “and also to help us find the appropriate elevation to put the cave.”

Computing clout

“When we are looking at a new project, we have to evaluate if we are going to mine it, or if we are not going to mine it,” says Gert van Hout, who oversaw recent development of the software program. “Within an hour, we can have results.”

“I can’t think of any program before CavePlanner where you could do that in such a rapid way, which is of particular importance at the order of magnitude or prefeasibility study level,” he says. “If you use common cave scheduling software, you have to assume the drawpoint spacing, you have to assume the direction of the layout, and you have to carve out the footprint perimeter. It could take a whole day.” CavePlanner output is used to guide the set-up of more rigorous simulations, from where to initiate the cave to how to undercut sequences and directions.

CavePlanner runs exclusively off the block model, which is generated from drill hole data. “It explores different mining sequences, very broadly, and allows us to investigate the optimum value sequence,” says Moss. “We may not select that because of other criteria that we have that would go against it – for example, it could show a geometry that is a bit too aggressive or optimistic – so we would use our judgement to reject that sequence.”

“It’s really a comparative tool,” says van Hout. “It gives us a rough idea of where we should be starting. It generates the tonnes of each of the metal grades per column, and gives us an indication of what values are sitting in there: what blocks will be economical and what blocks won’t be economical.”

Creation story

“Existing software didn’t address the comprehensive evaluation of all the possibilities for the sequencing of the ore extraction,” explains Glass. “The number of models that have to be produced is almost infinite, so the question was, could I develop a method in which the number of permutations that had to be evaluated was computationally manageable?”

Glass presented his prototype in May 2008. Further development was undertaken to address the needs of mine planners: improve on the prototype, make it more flexible and incorporate time-dependent properties. “The main difference between the two versions is that the first one was an Excel Visual Basic for Applications macro, while the current one is a stand-alone program featuring rapid calculation through parallel processing and enhanced functionality.”

The final program was completed in October 2009, and the hard work has paid off. A pilot test convinced Rio Tinto to change the mining sequence for one of its mines, resulting in hundreds of millions of dollars worth of added value. Rio Tinto has so far applied the program at four major projects.

“CavePlanner allows us to make better, more-informed decisions, and to do so at an earlier stage in the project,” says Moss. “This is a first look. It’s intended to be used at an order of magnitude level. It’s not a detailed planning tool; it’s a conceptual planning tool.”

Controlling variables

A number of geotechnical and mining constraints can be applied to CavePlanner before it is run, including the geometry of the undercut face, column heights and drawpoint maturity rules, as well as denoted production rate curves. Constraints can also include financial factors like net smelter return values, discount rates and operating and treatment costs.

The program can also account for variables that may change over time. “Maturity rules are very important,” says Glass, “where the draw rate from new drawpoints is gradually increasing with maturity up to a user-defined maximum that applies for the remainder of the production life of that drawpoint.”

“After you set the elevation and enter these parameters, the program goes off and simulates different undercutting sequences,” explains van Hout. “Then it’s a question of ranking according to net present value, and then it’s up to the mining engineer to select which ones make most sense.”

Net present value is perhaps the most useful piece of information at the initial planning stage. “At the point when CavePlanner is utilized, we are talking about an economic optimization, and what the most economically profitable route might be,” says Glass.

“There are a huge number of permutations,” he adds. “The precise number that it evaluates is very much dependent on the scale of the project. The larger the project, the larger the subset of models that CavePlanner produces.”

“The number of permutations really depends on the block model and on the undercut sequences that you want,” says van Hout. “If you have very stringent constraints, there are hundreds; if it is completely wide open, it is thousands.”

CavePlanner provides a solution to a problem Rio Tinto is encountering with more frequency. “Tonnage is getting bigger, and investment levels are getting huge,” explains Moss. “A big panel cave is a $10-billion investment, so we’re developing tools to help us better analyse them.”

“Design tools have always been critical to caving,” he says. “What we’re trying to do is to bring a bit more rigour to it.”

Related stories:

Post a comment


PDF Version