You wouldn’t blindly ask Siri or Alexa to make a life-or-death decision.
Get the weather, set reminders, or find directions, sure. But while the voice-activated intelligent digital assistants integrate well with modern life in some respects, most if not all must be supervised, at least to some degree.
The same goes for geotechnical monitoring technology, data acquisition, and resulting decisions. When lives are at stake and safety is number one — as it is across a mine site — an informed, considered directive must come from all available sources – marrying advanced technology and human experience for a safety-critical, optimum outcome.
In recent years we’ve seen ongoing, accelerating growth in the geotechnical monitoring technology sector. The mining industry has finally, it seems, begun to realise the true value of consistent, effective geotechnical monitoring strategies (both onsite and remote) for not only improving safety, but also for the remarkable and easily measurable impact on operational efficiency and profitability.
As the technology rapidly advances, however, so does the complexity of the tools and software, and the expectation remains that operational geotechnical engineers will somehow find the time to upskill to ‘expert level’ and very quickly learn how to use them effectively.
With complexity and capability comes high potential value-add, but also at high capital and operational cost. Hardware, software, commissioning, maintenance, technical support, training, data management, communications systems; the list goes on. And these things aren’t cheap, with management expecting a return on that investment. They also expect their very expensive geotechnical engineers will be able to facilitate that value-add.
To truly realise the full potential of these monitoring tools, sensors, and data, it takes many years of experience with real-life application. Similarly, it takes enormous amounts of data for a machine learning algorithm to come close to doing the same, and these tools typically still need to be supervised. This all must be balanced with the ever-increasing expectation to mine faster, safer, and more sustainably.
For the past decade I’ve had the relatively unique opportunity to travel the world training geotechnical engineers to analyse and interpret radar data for monitoring of open pit mine slope stability. At the same time, I was able to watch and learn how they do it themselves, usually self-taught, as I had also been previously when working in the field.
I learnt quite quickly while training that ultimately, the easiest way to help these engineers was just to do it all for them. But, the inevitable questions come in: “Surely for the price we pay for this technology, it should just be point and click by now?” and “What have they been doing all this time, developing this technology?”
Valid points, however, unfortunately, the technology isn’t quite there yet. There is still a human element required, and until the geotechnical monitoring data being collected daily by mine sites is shared openly with AI developers, it will be some time before robots can do it all for us.
Unfortunately, also, operational Geotechnical Engineers don’t typically have the opportunity to gain the breadth of experience or the depth of technical knowledge required to confidently analyse and interpret data in all scenarios, for all rock types, under all conditions.
And that’s just for one sensor type (radar for slope stability). Imagine finding the time to analyse, interpret and correlate data from hundreds of different sensors, all in real time, with full confidence to report on the output.
There is software that claims to do it but – call me a sceptic – it’s important to remember that the conclusions and advice from the site geotech engineer could and will at some point mean the difference between life and death.
So, what do we do? Wait and hope that robots will do it all for us? Soon, maybe. But not yet.
As we move towards a near future where the digital twin becomes closer to a reality in mining, it’s important to follow the old numerical monitoring adage: “rubbish in, rubbish out”. We need to do better when it comes to monitoring, analysis and interpretation but also collecting, cleansing and maintaining geotechnical monitoring data sets.
If we don’t improve our focus and attention towards developing robust geotechnical data monitoring and acquisition strategies, focusing particularly on long-term data acquisition with real-time, continuous data analytics and interpretation, with simple intuitive software tools to facilitate that, then we will continue to see the same catastrophic failures where “the monitoring instruments saw it, but no one was watching”.
At present, there are a handful of off-the-shelf commercially available multi-sensor aggregation platforms available, but it’s not a simple task to build an affordable enterprise solution for effective real-time, long-term deformation data analysis and interpretation with a truly customer-centric focus, it must be developed by the user for the user.
It’s always with hindsight, and regret knowing that surely, we could have done more, or done something differently, where we look back and say, “hey, that deformation trend was clear for months before the collapse”.
Too little, too late, because no one was watching, or the data just wasn’t clear without insurmountable processing work.
We have the knowledge, we have the (disseminated) tools, we have the people, but do we have the time? Do we have the money? Can we really afford to even suggest that we DON’T have either?
Real-time, long-term, continuous data analysis and interpretation is needed. The industry needs to move away from relying on the standard back analysis process, whereby a third party is engaged to “tell us what’s going on”, wading through a mountain of data that nobody’s looked at or maintained for years.
There are plenty of case studies to illustrate this flawed approach.
At an Australian open-cut mine recently, a tailings dam failed and the InSAR (satellite radar) data was very clear. It showed for about three months beforehand that the structure was moving. Something was changing.
Fortunately, the result was not catastrophic in that case and the response was managed well.
In the well-known fatal Mariana dam disaster in Brazil, 19 people died in 2015, and two mining giants were eventually fined billions following the deaths and the environmental devastation that resulted.
In learning from these, and other instances, it has never been more apparent that we need to get ahead and place more emphasis on the importance of long-term mine monitoring.
But it’s not just about relying solely on technology, with blind faith. Often monitoring tech can be used in the wrong way, and this actually slows things down – when alarms go off, nobody knows why, and work is stopped – often without needing to.
There’s considerable merit in engaging with support to be able to do it properly. Improving safety, along with operational efficiency and profitability, are key value propositions for any technology these days.
When the lights are all on – and the information is all there – it’s vital that somebody’s home.
Peter Saunders is a Principal Engineer at Cartledge Mining & Geotechnics, bringing over 21 years of experience in geotechnical engineering and technological innovation. His expertise drives industry best practices by offering independent advice on risk management strategies for large open-pit mines, slope monitoring techniques, business development, and project management. Peter pioneered a 24/7 remote monitoring service that supports hundreds of slope stability radar users worldwide.
Images: Cartledge Mining and Geotechnics
