Mining is all about profitability. Revenue versus cost. But tracking that can prove tricky, particularly when you're relying on people on the ground to report on activities. It all comes down to big data: the more data you can collect, the more information you have at your disposal, and the more insight you have into the business.
IOT technology is paving the way for mine optimisation initiatives, says Johan Pietersen, MD of Virtualscape Technologies. "Mines are looking for the ability to gather data for various purposes, including improvements in efficiency and safety. They need to be able to monitor activities, assets and people and make better decisions based on real-time data. AI applied to such data is able to provide a realistic view of the mining operation's performance on an hourly basis, allowing mine managers to make pro-active judgment calls on operational execution to their financial benefit."
Pietersen continues: "We were approached by open-cast mining companies who wanted to be able to collect certain data that would enable the mining operation to run at optimal efficiency by gaining insight into the revenue and cost elements of the mining process."
The initial challenge posed was tracking how much soil was being moved from the pit. Operations are set specific targets in terms of bulk cubic meters (BCMs) of soil moved over predefined periods. Traditionally, each load is recorded manually by someone on-site. The problem with this is two-fold – usually the task is assigned to a group of people, who also have other tasks to complete on the mine, so it may not be an accurate record; and the figures have to be shared so that they can be recorded and processed, which is usually accompanied by delays and limitations.
The challenge is that the inaccurate data on operational activities can lead to penalties or over-expenditure during the month, only to be accounted for at the end of every month, which is a cost or revenue issue for all involved. The veracity of this is often in doubt owing to inaccurate figures.
Pietersen says: "The mines needed an independent, objective way of monitoring performance on site. We started out by simply being able to count the loads of earth being moved. From there we branched out, creating the ability to convert the loads to different metrics such as tons and BCM based on the types of material being moved, the volumetric capacity of the vehicle being used to move it and material densities.
"Following on that, was the ability to work out how many of the loads moved were ore or not. Then we automated the ability to monitor the number of hours that each machine is active/productive per shift." He explains: "The machine hours consumed directly relate to cost. The client was able to calculate how productive each machine was being on site versus time spent stationary."
The third part of the challenge posed was to control who could access specific equipment on the mine based on licences held. Pietersen explains: "The mining industry is very strict about operators only operating equipment for which they are licensed. This is generally referred to as key control."
Looking to the next step in evolving the solution, new requests include the ability to generate daily costing sheets for the equipment themselves so that they can work out how much each machine's running costs are per hour, per trip or per shift, compared to revenue generated.
The advantage of automating the generation of all of the above-mentioned information, says Pietersen, is that where previously the mine would have to wait for a shift to end before finding out how many loads were completed, automation means that this can be monitored live with high precision.
"The ability to monitor the site on an hourly basis means that if rain delays production, for example, we're able to calculate exactly how many additional vehicles need to be deployed to site to make up the lagging loads to still make targets."
Automation also improves safety on the mine. Over and above allowing only authorised licensed operators to access equipment, in the current roadmap operators will be able to flag hazards on site that could impact productivity or safety so that other operators are warned and technical teams can take action, if required.
Pietersen says: "We're in the process of building a refuelling dashboard that will remotely monitor fuel levels and prompt operational control staff to refuel equipment when necessary to ensure continued production and avoid bulk refuels or breaks in production. A machine that runs out of fuel on site is a safety hazard and impacts on productivity.
"We'll also be able to proactively identify operator fatigue by monitoring the patterns of the operator's shift, comparing downtime to productive time and combining that information with data from fatigue monitors located inside the vehicles. If the operator seems unwell or fatigued, action can pro-actively be taken before productivity is impacted, in addition to safety concerns. AI excels specifically at combining and processing large volumes of data for this type of application."
In addition to being able to track the vehicle's whereabouts on site, it's also possible to create rules around speed limits depending on whether the vehicle is under load or not. Automated driver alerts can be generated to slow down a speeding driver or inform him of potential hazards on route for example which leads to savings on maintenance and tyres, for instance. Pietersen says: "While a lot of the capabilities that we're developing speak to safety, they also optimise profitability for the clients."
"These are real-world applications of IOT and AI to improve the economics of the operation while significantly reducing the risk for people on site."
Looking to the future of AI in mining, Pietersen says: "It's already possible to adjust production to meet very specific cost and revenue targets based on data gathered and analysed in real-time. The next step will be determining where to mine more accurately. We're working on a blast advancement technology that uses markers to track the ore body as it shifts following blasting for more profitable mining.
Explaining how IOT is implemented in an environment where the average person has a vehicle and a cellphone, Pietersen says: "Sensors and gateways are installed on equipment and strategic locations such as dumps and along the haul routes. Data received from equipment sensors, movement and behaviour that pass these points are uploaded to the cloud via the gateways."
He says: "Interestingly not all of the technologies in their isolation are revolutionary or new, but the application of these technologies in our applications and combination is what brings the value. IOT is not just about connected devices nor is AI just about machine learning, it's about the application of those devices and the intelligence you derive from it to add real-world value to users."