Project concepts are often born out of a need to solve longstanding problems in exploration, metallurgy or mine planning.
Increasingly, however, they are now addressing issues that come with the advent of big data and envisaging the next generation of products that can add value to these datasets.
And of course, industry partners are essential for the subsequent development and testing stages before commercialisation is plausible.
Here are four recent CSIRO innovations — from concept to their current commercialisation stage.
Geophysics Processing Toolkit (GPT)
The Geophysics Processing Toolkit (GPT) is a Cloud-based and Web-serviced program that provides interactive visualisations of large datasets, together with suites of tools that assist inspection, processing and modelling of geophysical survey data.
Currently GPT is in the concept and technology evaluation stage and will initially be designed for electromagnetic (EM) data, which are large in volume, dense in time and space, and have many dimensions. EM surveys are widely used in mineral exploration, groundwater resource detection, soil mapping and geotechnical hazard assessment.
Research Director Dr Sandra Occhipinti remarks that CSIRO has a long history of involvement with EM, including the development of the Tempest system, SiroTEM, LandTEM and SQUID sensors for EM applications.
"Our team has strong EM expertise and several processing and inversion codes developed over the last 30 years," says Dr Occhipinti.
"GPT will provide a streamlined workflow for geoscientists to follow and to more easily load, visualize, process, invert and export the data. Commercial geophysical processing software is quite limited for EM data, so we are building a Web-based, Cloud-processing solution that is much more accessible."
A typical EM survey contains hundreds of thousands of individual data locations, with millions of measurements. Before this data can be inverted for conductivity–depth information, it must undergo a laborious QA/QC process to remove the noisy and inferior data.
GPT is envisioned to greatly accelerate this process.
The time is also ripe for GPT’s development — the entire Australian continent will soon be covered by airborne EM data through an acquisition program by Geoscience Australia and State and Territory geological surveys. It will be the largest EM survey in the world.
Initially a platform for EM data, GPT will also incorporate existing tools developed by CSIRO's Deep Earth Imaging Future Science Platform.
‘"We envisage that within two years, GPT will have grown beyond an EM toolbox and will incorporate other geophysical techniques such as gravity and magnetics."
"Testing of the platform will be needed and so we are currently interested in talking to potential service providers," says Dr Occhipinti.
Rosetta is a software platform designed for exploration and mining companies to help understand and predict orebody characteristics. It pulls together analytical data from a range of sources and provides predictions of ore quality for processing evaluation and mine planning.
Rosetta is in the early prototype and development stage, building on a proof-of concept stage that commenced in 2016 on the Olympic Dam deposit in a collaboration between CSIRO and BHP.
"During this next stage we plan to build a portfolio of case studies to further test and develop Rosetta. We want to be open with what we are doing and not have Rosetta seen as a 'black box' where data goes in and predictions come out," says Research Group Leader Dr Yulia Uvarova.
The Olympic Dam project was born out of a need to process and incorporate vast amounts of hyperspectral data created during exploration and mining. It was also at a time when machine learning and AI possibilities were being explored by the mining industry.
Dr Uvarova explains, ‘"Partial mineralogy was being collected by hyperspectral scanning, but we wanted to see what value we could add to these datasets using data analytics and machine learning."
"Was it possible to predict properties such as lithology, chemistry and mineralogy that were being generated by traditional laboratory-based analyses?" asks Dr Uvarova.
Using a supervised machine-learning approach to predict the properties was a perfect test case and proved to be very successful. However, the uncertainties between predicted and actual values were high.
‘"Since then, machine learning has significantly progressed and so by refreshing the original code with a few improvements we have shown that the results are now actually much better," says Dr Uvarova.
"Rosetta technology has proved that it works as a robust predictive tool. We now need to use large datasets from different ore bodies to increase the confidence in the predictions made, and we are looking for companies to work with to help achieve this".
Aimed at research across the geological, biological, environmental sciences, as well as material science and medicine, the Maia Mapper is a high-throughput X-ray fluorescence (XRF) detector system that produces high-definition, quantitative elemental images with microscopic or nanoscopic detail in real-time.
The technology was first developed in 2014 for use on the synchrotron, but the need to analyse large samples drove development of the laboratory-based Maia Mapper. After several years in the proof-of-concept stage, the Maia Mapper has now progressed to the prototype and development stage.
"Although other techniques, such as laser ablation systems may have similar or even better detection limits, the Maia Mapper's real strength is its combination of good detection limits, fine resolution and its ability to analyse samples over a large spatial extent, for example in drillcore up to 50 cm in length," says Research Group Leader Dr Mark Pearce.
"It fills a well-needed gap between the much broader sampling of drillcore that mining and exploration companies carryout, and the very detailed microanalysis of small samples that researchers carryout using lasers and microprobes."
The Maia Mapper fits neatly within CSIRO's drillcore laboratory facility, which includes the Minalyze XRF line scanner and the HyLogger spectral scanner.
"We are able to scan many hundreds of meters of drillcore, and then use that data to move down a scale and determine which 50 cm samples are representative of the drilled intervals and analyse them using the Maia Mapper.
"Once we've done that, we can move down a further scale and cut thin sections and carryout micro-analyses," says Dr Pearce.
Since early development, Maia Mapper has increasingly been used in CSIRO projects and is now also being incorporated into research projects for mining companies.
"We've recently been involved in projects looking at the geometallurgy of gold, that is, where the gold sits in the samples, if it’s hosted within sulfides or if its free gold that's going to be easily extracted.
This kind of data is extremely useful at a very early stage of exploration by helping determine how the gold will be extracted during later mining and processing".
detectORE™ is one of CSIRO's most recent success stories of technology commercialisation for the minerals industry. It was developed primarily to help gold explorers find deposits, faster.
The detectORE™ system comprises sample preparation, sample processing and use of a portable X-ray fluorescence device to obtain low-level gold results in the field. Commercial partner Portable PPB Pty Ltd won the technology license contract in 2019 to use and further develop detectORE™.
CSIRO Principal Research Scientist Mel Linton invented the detectORE™ technology and went on to become Portable PPB's Chief Technical Officer after leaving CSIRO in 2018.
Portable PPB now employs up to 19 people at various times and is working towards full commercialisation by the end of 2021.
The ability to analyse gold in the field has great advantages for exploration companies, particularly cost and time savings.
"detectORE™ is changing the way that traditional drilling campaigns are carried out, allowing companies to adapt on-the-fly and modify exploration plans and activities while still in the field. Cost savings are also substantial as the number of traditional laboratory-based analyses required is significantly reduced," says CSIRO’s Business Development and Global Manager Wayne Robertson.
Results to date are extremely encouraging. Managing Director of Portable PPB Simon Bolster has been working with several companies in the Kalgoorlie region.
"We tested detectORE™ over a large area of interest and confirmed anomalies identified by traditional sampling and drilling. We also found an additional area that fell between drillholes that they weren't aware of.
"By reducing the sample grid, detectORE™ then identified a very coherent gold anomaly, which was later confirmed by laboratory assays that came back after nine weeks," says Bolster.
Being able to analyse soils and drill chips for gold in the field can lead to the identification of deposits faster and at lower cost.
detectORE™ has a quick assay turnaround of just eight hours, compared with potentially weeks from a laboratory, and most significantly can identify gold down to less than 20 parts per billion (ppb).
Being able to rapidly identify areas with higher gold levels compared with low background levels without sending samples to the laboratory, means that real-time decisions for soil sampling and drilling campaigns can be made during the field exploration campaign.