Capitalising on your MWD data

WHEN it comes to combining different monitoring data, it appears the sum is greater than the individual parts. Sarah Belfield writes

EXTRA productivity gains could be wrung out by marrying measurement-while-drilling data, haul-truck traffic data and fragmentation imaging data, according to a recent study of a Scandinavian copper mine.

The study, which looked at the Aitik open-pit copper mine in Sweden, was completed in June by Shahram Mozaffari at the Luleå University of Technology.

Mozaffari’s study analysed selected Aitik benches to assess the potential for combining measurement-while-drilling (MWD) data and image-based fragmentation data to predict rock conditions and so improve blast design and productivity.

Mozaffari said while both data types were saved to database at the mine, to date they had not been fully used.

The Aitik operation annually produces around 18 million tonnes of ore. An expansion worth 5.2 billion Swedish krona in capital expenditure has been underway there since the mine’s owner approved the plan in October. The eventually allow mining of 36 million tonnes a year. The ramp-up in operations was anticipated to begin in 2010.

As at the start of 2007, proven reserves stood at 526 million tonnes grading 0.28% copper, according to the mine’s owner Boliden, a Swedish base and precious metals company involved in mining and smelting. There were also 99 million tonnes of the same copper grade in probable reserves. Aitik’s measured and indicated resources came to 858 million tonnes. The ore body also contains gold and silver.

MWD has been successfully used to optimise blasting in open-pit mining generally, according to Mozaffari.

MWD has been in use at Aitik since 1998, he said. Data routinely collected at Aitik during blasthole drilling included depth, time, rotation speed, the weight on the bit, torque, vibration, cuttings-removal air pressure, and penetration rate. Other figures, which are calculated, are recorded as well, such as blastability index, comminution index, and specific energy (per-unit work). The Aquila DM-5 drilling management system was currently in place at the mine to carry out MWD functions.

All up, eight distinct rock types could be found at the mine. The smooth transitions among similar rock types, such as biotite schist and biotite gneiss, usually made it difficult to pinpoint lithological boundaries, Mozaffari said.

Bench 4141, which featured four of those rock types plus detailed information from six exploration drill holes, was picked for close study.

Mozaffari found that when contour maps of the bench area were plotted up using MWD data from blastholes, the most promising map was the one depicting the variation in average downhole penetration rate.

“Rate of penetration provides the most accurate information about the geological situation at Aitik,” he said.

It meant at this particular mine this kind of map was a good proxy for the variation in rock strength. Running through the bench studied was pegmatite — generally notorious for its great hardness — and the penetration rate contour map clearly picked it out.

And while the contour map of torque also showed merit, he found similar maps involving specific energy, blastability index, or vibration weren’t as useful, perhaps because of noise present in the data.

Mozaffari then looked at the mine’s image analysis data, which stored details about fragmentation size distributions. According to Mozaffari, Aitik had been using the Split-Online fragmentation analysis system since 2003.

Because Aitik was also using Caterpillar’s MineStar truck monitoring system, it was possible to track every truck load from its starting coordinates on the bench to the dumping point at the primary crusher. This made it possible to marry up fragmentation image analysis data with MWD penetration rate data from the corresponding blastholes.

“[F]or the selected bench there is a significant correlation between registered penetration rate and monitored rock fragmentation,” he said. As one might expect, fragments tended to be smaller when blasthole penetration rates where higher.

But the power lying dormant in the separate data was indicated by an example graph Mozaffari produced.

By regressing fragment size data on penetration rate data across the spectrum of sieve size fractions (0% passing to 100% passing), he came up with a fragmentation curve for the specific penetration rate of 60 centimetres per minute.

The implication was that the curve could be shifted and stretched in the desired directions based on an iterative, learn-as-you-go approach to blast designs.

“This study shows that it should be possible to proactively use both MWD data and image analysis to achieve more uniform fragmentation by adapting … charging and blasting design,” Mozaffari said.

While he cautioned that further work was needed on more Aitik benches and larger areas of the open pit to confirm the findings, he recommended that the process of linking image analysis data with haul-truck load data be automated.

He also recommended that a method be developed to better integrate MWD data into blast design optimisation processes to help achieve more uniform fragmentation.

N Boliden AB

+46 8 610 1500

To keep up to date with Australian Mining, subscribe to our free email newsletters delivered straight to your inbox. Click here.