The extraction industry is turning to the Internet of Things to squeeze more out of the ground.
Across the globe the productivity of mining and mineral extraction has declined by 20 per cent over the past seven years, despite the push for increased output. With declining market conditions, productivity differences between the best- and worst-performing mines are stark, with some of the best-practice outputs coming in at more than double that of the median performers.
Speaking at the Melbourne Mining Club in February, Alan Davies, chief executive for diamonds and minerals at Rio Tinto, said: 'We have seen iron ore prices steadily decline, oil prices dramatically slump, copper and aluminium moving in all manner of directions. However, the long-term fundamentals are pretty clear. The world will continue to demand the metals and minerals that make modern life work. By 2050 the world's population will increase by a third to almost 10 billion people.'
From an ICT perspective, mining is an industry which is viewed as risk-averse. Underground activities in particular have remained resolutely manual, seemingly sitting years behind the technology adoption curve. The vast majority of operations use antiquated communication systems that would be unable to support modern applications such as autonomous vehicle operation, ventilation-on-demand and real-time location. Yet in a few cases, owners have taken a hard look to determine how modern wireless technologies in underground and surface operations can provide greater insights into their business.
They are implementing industrial systems that can adjust to their own environments or even their own health. Instead of running until failure, machines schedule their own maintenance or, better yet, adjust their control algorithms dynamically to compensate for a worn part and then communicate that data to other machines and the people who rely on those machines. By making machines smarter through local processing and communication, the Industrial Internet of Things (IIoT) could solve problems in ways that were previously inconceivable. But, as the saying goes, 'If it was easy, everyone would be doing it.'
The IIoT challenge becomes even more daunting and complex when comparing the requirements of the industrial Internet to those of the consumer Internet. Both involve connecting devices and systems all across the globe, but the IIoT adds stricter requirements to its local networks for latency, determinism, and bandwidth. When dealing with precision machines that can fail if timing is off by a millisecond, adhering to strict requirements becomes pivotal to the health and safety of the machine operators, the machines and the business. As a result, take-up of IIoT in safety-critical industries has been slow and conservative.
The mining industry has for a long time used satellite communications to monitor wells, But with the increased infrastructure of cell towers and decreased cost of wireless devices and data packages, it is much easier today to pass across data captured by sensors in remote areas using terrestrial networks and on into the Internet.
Connected sensors on instrumentation provide engineers with real-time visibility into ground movement; networked drills provide feedback on drill times, operator behaviour and drill performance. Connecting production vehicles and an on-board network of sensors to reporting, analytics and decision support allows mine operators to identify inefficiencies, anticipate equipment failure and deliver an overview of the entire mining process.
Chile is home to some of the deepest open-pit mines in the world. Owned by Codelco, the world's largest copper producer, the Chuquicamata copper mine is situated 1650km north of Santiago, is 4.3km long, 3km wide and more than 850m deep. The company's future is closely linked to the success of a massive project to transform the pit into an underground mine, enabling the exploitation of copper and molybdenum for an additional 40 years from 2019. Greenfield projects such as this are now almost exclusively designed to accommodate the IIoT.
However, more mining operations are also deploying high-speed networks to connect their disparate legacy array of sensors to remote monitoring systems that feed valuable data to decision support systems. Kazakhstan's largest producer of uranium, KazAtomProm, has through its Green Smart Mine project with UK IIoT specialist Intellisense.io installed wireless mesh local networks and a long-range wireless backhaul network to coordinate the collection of data from multiple pumps that are used to liquefy ore and bring it to the surface from where the heavy metal can be extracted.
The Brains.app software developed by Intellisense.io, a web-based analytics application, delivers recommendations to the mine operators, gains insight into the efficiency of the well-pumping system and tracks mine performance over time. Serik Kozhakhmetov, CEO of KazAtomProm's Institute for High Technologies, says the companies are now partnering to 'roll out the technology to mining sites across Kazakhstan'.
Miners in Chile have relatively long experience of using on-board sensors for condition monitoring: since April 2008, nine electromechanical shovels for open-pit mining have been continuously monitored at four different open-pit mining locations in Chile, including two of the largest copper mines.
In addition to continuous monitoring systems, Cadetech, a'systems integrator of National Instruments equipment, developed several portable instruments for shovels. One of these is based on the NI CompactDAQ system and the other uses the CompactRIO and the NI TPC-2006 touch-panel computer to configure a fully autonomous, rugged 16-channel instrument in a suitcase form that performs vibration analysis.
Electromechanical shovels for open-pit mining are huge, mobile, non-stationary machines used to load haul trucks, which transport ore to processing plants. Usually, the shovel-to-trucks ratio is about 1:12, so unexpected shovel downtime has a pronounced effect on production.
Traditionally, it has been difficult to apply condition monitoring and predictive techniques to the shovels, due to inadequate analysis algorithms and equipment, as well as the harsh environment. Traditional vibration analysis - the main tool for predictive maintenance on rotating machines - performed by conventional equipment is based on the Fourier transform, which assumes constant rotational speed. This is not adequate for the shovel, therefore a different approach was needed.
The urgent need to move to a predictive, proactive maintenance strategy led to the creation of SiAMFlex, an Advanced System for Flexible Monitoring. SiAMFlex was at first an initiative of Professor Pedro Saavedra at the University of Concepcion, Chile. It began with research to develop a vibration analysis algorithm suitable for the electromechanical shovels. Once the algorithm was ready, the second stage was to implement this technique as the core of a continuous monitoring system. Now, SiAMFlex is supported and continually updated by Cadetech to maintain a complete mechanical and structural asset integrity management and analysis tool based on CompactRIO.
'On the shovel, an on-board CompactRIO system acquires simultaneous signals from accelerometers, encoders, and strain gauges,' says Cadetech's Daniel Ramirez. 'Vibration and strain signals are continually monitored and compared to alert and alarm set points as a first indicator of trouble. Signals are periodically stored at user-defined intervals in case of an alert or alarm.'
In this case, the monitoring application on the CompactRIO system searches for the best measurement periods to analyse and optimise the signal-to-noise ratio. With this approach, data is stored at regular, predefined intervals to control eventual mechanical changes in the machine, and data is recorded when a sudden event occurs. In both cases, complementary signals from the shovel control system are stored for reference and to enhance the possibilities of proactive corrections.
Mine of the future
Rio Tinto's 'Mine of the Future' aims to go beyond maintenance prediction and use sensors to control the vehicles that move around its mining operations, such as the iron-ore mines in Pilbara, Australia. Rio Tinto's operation in Pilbara includes an integrated network of 15 iron-ore mines, four independent port terminals and a 1,700km rail network. The company sees automation as the main driving force for improving efficiency, with operations guided from a facility in Perth, more than 1,000km to the south. Rio Tinto is set to become the world's largest owner and operator of autonomous haulage system trucks. There are currently over 50 autonomous trucks in operation at the Pilbara sites and the number will grow in coming years. The company has also worked on automated drilling systems, trialling one in 2008 at the West Angelas mine in preparation for deployment across the Pilbara operations. Automated trains will move the ore out of Pilbara; the company is spending more than half a billion dollars on the programme.
Rio Tinto is also working with 'big data', identifying patterns to allow it to predict and enhance performance within its operations. 'While we are garnering significant benefits from our technology investments, I think at an industry level technology can still do so much more for us,' says Davies. 'The next leap we are now starting to see is how technology can speed up learning. Traditionally, if we wanted two operations on opposite sides of the planet to learn from each other, that was a logistically messy proposition with an uncertain payoff. People flying all over the place.
'But a lot changes when you see those geographically distant operations as sources of data - big, real-time and proprietary data. For example, in Brisbane, we have what we call our Processing Excellence Centre. In that centre, we bring together live data-streams coming off the mill circuits and float tanks of the copper concentrators in Mongolia and in Utah. We can see and compare everything that's driving performance up or down. We can have sites on opposite sides of the Earth learning from each other in real time.'
The push for efficiency has seen some of the precursor technologies to IIoT being repurposed. Oil prospectors have found a way to use the RFID tags originally developed to track goods as they move around a distribution network. Hardened for extreme environments, the tags can be used to open and close the wells at precise points.
In modern oil drilling, the borehole is lined with casing elements that are designed to stop water getting into the hole and mixing with the oil. To access the oil, drillers use perforating guns fired into the side of the pipe at the predicted depth of the oil reservoirs. The problem is firing the gun at the right depth. Marathon Oil came up with the idea using RFID tags for the job over a decade ago. In this method, tags are placed in the casing elements, so that they can be identified individually once placed inside the borehole. A scanner fitted to the perforating gun as it descends reads each tag in turn, firing when it has passed one known to be at the depth of the oil reservoir.
Offshore drilling provides another environment where highly connected smart devices can aid automation in areas that present a danger to humans. They also help to scale down the equipment needed to perform surveys and drilling. Traditional drill ships are very large and expensive to construct and maintain. Although this scale is required for some operations, in others, such as sampling, testing and geotechnical investigations, a drilling system carried by a remotely operated vehicle (ROV) can be much more cost-effective.
Canyon Offshore designs and develops drilling using a drill mechanism originally created by Cellula Robotics, its control software optimised by Metis Automation. Legacy implementations of this type of drilling system in the industry depend on extensive manual control, requiring an operator to be fully trained on what the system can and cannot do - the large forces involved mean the drill can easily be damaged. The operator would have to control the state of individual valves, rather than focusing on the high-level drilling task at hand. Ultimately the result was that operations took a long time to complete and there was a risk of human error. For Canyon Offshore, the goal was to create a platform that offered consistent, repeatable drilling operations. And to achieve this, automation was key.
The ROV drill consists of two main components: the subsea drill system and the topside, above water, control system. The drilling module is surrounded by tool racks with a robotic arm used to pick up tools and move them to the drill head. There are also various motors, pumps and other components that allow the drill tool itself to operate.
In the control room on the deck of the ship, the ROV pilot sits at a bank of monitors. They are presented with information to control all of the complex drilling processes. They have a couple of touch-screen monitors that are running software applications and camera feeds to visualise what is going on subsea. The pilot has a full-featured chair with joysticks to give full control over the whole system. The system is designed to make the control system as familiar to ROV operators as possible, while including the advanced features that drilling requires. Going from job to job the drill can take on very different challenging operations, the goal is to give the drillers as much control over meeting these requirements as possible in order to complete successful drilling operations. Both of these components, topside and subsea, work together.
'To achieve the level of automation that was required to make the system successful, installed in the subsea structure there is a CompactRIO embedded device from National Instruments, running an application written in LabView,' says Tristan Jones, responsible for technical marketing at National Instruments. 'It is connected over Ethernet, via an umbilical, to the control cabin on deck. Topside there is a PC that is running another LabView application that is in constant communication with the subsea system. This configuration is ideal because the functionality of the system can be segmented in order to have all of the safety critical systems and tight control loops running on the CompactRIO embedded system subsea.
'For the user interface, the goal was that if subsea visibility is so poor that you cannot see the operation of the system, the data and functionality that the control system provides will effectively allow the operator to safely use the system 'blind'.'
The system contains lots of tool configurations pre-saved, so from job to job, it is possible to load in the appropriate configuration. This significantly cuts down the time that it takes for the driller to configure the system for a new drill operation. The system can even be reprogrammed remotely to avoid damage.
In one project off the coast of Africa, there was reported water ingress into one of the non-critical interlock sensors. Pulling the drill out for repair would have taken at least 24 hours. However, due to the ability to access the system from onshore, Metis Automation was able to develop revised firmware for the system and deploy it to the embedded system subsea, meaning that the drilling operation could continue safely and without significant delay.
'Now, a client who is monitoring the operation of the system may say 'it would be great if we could do this', if the drill does not support it, this can be added as a real-time software update while at sea,' says Jones.
By supporting real-time software updates, mining operations can be upgraded almost on a daily basis although companies will need to take care their systems are not hacked. Always conscious of physical security, companies will need to invest in hardening their IIoT networks as they evolve. But they will be able to minimise downtime and extract minerals using less energy and materials than before.