AI predictive maintenance is the holy grail for mining
We interview Mark Majzner, General Manager at Razor Labs, on the biggest challenges facing the sector right now and the importance of the right technology.
23 February 2021
Mining Innovation Director
There is no substitute for going out on site and speaking to the people with red dust under their fingernails. As an AI partner working with large mining organisations, our two greatest tools are a high-vis shirt and a pair of steel toe cap boots. It is only through intimate relationships like this that effective innovation can be achieved, and we at Razor Labs have been working closely with our partners in the mining space throughout the COVID-19 pandemic.
The ongoing crisis has not left the mining industry unscathed, but it has been an accelerant for AI adoption that will be crucial for meeting the demands of the future. We have seen great stability in the past year among early adopters of AI and remote operations solutions, particularly those that have taken an integrated, thoughtful approach to data. By making the right decisions for AI adoption now and targeting the right applications, mining organisations are unlocking immediate value during these challenging times while building strong foundations for the future.
How do you believe the pandemic has impacted the mining sector in particular? What do you see as the biggest challenges for the sector right now?
We have actually seen innovative mining companies hit production records during this time, while slower adopters have had to close down in some cases.
In the first six months of the pandemic, mining companies were understandably focussed on ensuring safety and maintaining production. During this period the value of remote operations quickly became apparent across the industry for those that had invested in it. Here in Perth a number of mining organisations were able to remotely drive their trains, operate reclaimers, load the ships and make critical decisions, proving vital for business continuity in a crisis.
Until the arrival of COVID-19, the progress toward reducing numbers of people - from remote and potentially hazardous environments - was relatively slow, but the arrival of the pandemic allowed miners to quantify the true benefits of remote operations. We have actually seen innovative mining companies hit production records during this time, while slower adopters have had to close down in some cases.
In reality, many organisations are just starting out on their AI innovation journeys, perhaps using some standalone applications like autonomous trucks, driverless trains, LiDAR (Light Detection and Ranging) on drones and video analytics. While this is progress, the most valuable thing these organisations can do now is to improve the collection, hygiene and storage of data to enable more integrated solutions going forward.
What do you believe companies should be doing in order to bounce back and thrive again? How important do you think technology is?
A real pain point being felt by one of our customers involved car dumpers, these are big, expensive pieces of equipment that are essential to production. It is here that we were able to support them with AI-based predictive maintenance and increase asset utilisation by 20%, which meant the solution paid for itself in under two weeks.
Technology will be of vital importance for mining companies in bouncing back from recent challenges, with a sharp focus on safety, cost and production. Video analytics and collision prevention systems are some of the easiest options for companies to instantaneously improve safety and overall efficiency. Moving a step beyond this, we are seeing immense value in areas like malfunction prediction and asset health monitoring, via which organisations can maximise production by planning their downtime.
We at Razor Labs have enabled large mining organisations to significantly improve the efficiency of existing assets through malfunction prediction technology, directly benefitting the bottom line. To achieve this, it will be important to invest in three key areas. The first is IIoT (Industrial Internet of Things), followed by investment in data capture and storage, as well as gaining an understanding of the change management systems involved in properly cementing new systems into an organisation.
How can AI help mining companies maximise productivity and run operations more efficiently?
It is worth reiterating here that the process of adopting AI is a journey, and truly maximising productivity is a goal that has to be achieved with gradual progress. A common place for mining organisations to begin this journey is with decision support, rather than diving straight into closed-loop, autonomous systems for optimising things like mills and crushers. Often mining companies want to start out by having us alert them to things like maintenance requirements, to build trust in the systems and then take the necessary action themselves. As data quality and relationships develop, solutions can become more integrated and effective.
Our mining partners always reinforce to us the importance of creating AI solutions that are robust, accurate, useable on site and scalable across their organisation. This is in Razor Labs’ DNA and influences how we approach and design all our AI projects.
A real pain point being felt by one of our customers involved car dumpers, these are big, expensive pieces of equipment that are essential to production. It is here that we were able to support them with AI-based predictive maintenance and increase asset utilisation by 20%, which meant the solution paid for itself in under two weeks. Another example is the success we have had in using LiDAR data to detect, classify and create alerts for changes in underground mines, fragmentation sizes and ore humidity levels.
How do you see the mining industry adopting AI? Are there any areas which are better than others?
There are still very few large mining companies working with AI in a structured way. Machine learning is being used to a greater extent, but very few are using deep learning effectively. There may be two or three making significant progress with AI, with different industry segments proving very siloed. For example, companies focussing on copper may well take a completely different approach to those targeting iron ore.
I have touched on the importance of developing an AI strategy including IIoT, and capturing and storing data for effective AI adoption, but another key factor is the way mining organisations and AI partners work together. I believe that great innovation can never just come from head office, the genesis of great projects has to come from working closely with the mining engineers working on site. Engineers on site best understand the problems and AI partners have a deep understanding of where AI is most valuable according to the company’s innovation risk profile.
Additionally, AI partners bring experience with other customers in other sectors and industries and employ the very best and brightest AI developers who can solve the most complex problems. It is only through doing this that an accurate understanding of the challenges can be gained, and the best next steps taken towards reaching a solution.
What other digital innovation trends have you seen across mining and other sectors? Which do you think are here to stay?
My view is that in the short term there will be a much wider spread adoption of video analytics and LiDAR technologies as mining companies recognise the immediate value they can gain. If your mining organisation has already invested in cameras and installing LiDAR, the opportunities to generate data insights are almost unlimited. The next level of benefits will come when organisations continue to invest in their smart IIoT strategies and to leverage its value by adopting AI for things like malfunction prediction and process optimisation.
While this prediction doesn’t focus on other big trends, it frames the importance of proper implementation that will actually lead to results. We are actively supporting our clients along the AI adoption pathway, helping them make the steady progress that prevents projects getting stuck in the ‘proof of concept’ valley of death.
This article is a part of our Innovation Leaders in efficiency series. To view the report and further interviews and insights into efficiency solutions, visit our Innovation Leaders page here.
Learn more, explore our insights and reports
Mining’s robotic future
As mines will inevitably be more remote, much deeper and riskier, how can automation help?
21 September 2020
Reducing unplanned mining plant downtime through AI-based predictive maintenance
Metals & Mining
Unplanned downtime is a major issue for throughput. This advanced AI-based predictive maintenance solution can reduce failures, lost production, spare parts use, labour costs, whilst increasing throughput.
Return on Investment
What industry-specific challenges are you facing right now? Talk to us so we can find proven solution to help you.