Insight
Improve mining ’ s artificial fitness
Using the analogy of personal fitness tracking tools , we can better visualise the potential of artificial intelligence ( AI ), Internet of things ( IoT ), and edge computing in the mining sector , writes Louise Steenekamp .
A
simple analogy in the everyday realm of consumer technology can be a powerful way of understanding how advanced technologies like AI , the IoT , and edge computing can advance the mining industry .
Think of the wearable fitness tracker — which burst onto the scene a few years ago — now being used by sports and health enthusiasts across the world . We can draw some interesting parallels between these personal devices and advanced industrial mining technologies .
The beauty of your smartwatch is that it stays connected to the rhythms of your body , capturing data relating to step count , sleep , heart rates , activity levels , energy burnt , and even body fat . When connected to the Internet , it pushes all this information into databases , to be analysed with sophisticated tools .
In just the same way , advanced numbercrunching algorithms in mining technology ingest real-time data from a variety of equipment sensors — tracking vibration , temperature , wear and tear , running times , and more .
Data is converted into actionable insights , giving ‘ definition ’ to the data . As mentioned above , the raw data from personal health devices is shaped into forms and definitions that we can understand . With the smart use of AI systems , the general concept of ‘ motion ’ becomes the tangible category of ‘ steps taken ’; ‘ activity ’ is converted into ‘ calories burnt ’; and so on .
AI enables us to analyse the information detected by accelerometers in a smartwatch or fitness device , interpreting and calculating it into the statistics that we can understand .
In the same way that our fitness gadgets start to sort and classify data on the device , edge computing makes it possible for much of the data analysis and AI services to happen on mining devices themselves — reducing the vast volumes of data that need to be sent back to central processing hubs . With edge computing , we can process and translate data into actionable insights , so that only relevant information is transmitted to , among others , central systems and on-board computers .
The whole purpose of fitness tracking devices is to stimulate a change in our behaviour : to exercise more , to eat healthier , or to get more sleep . We set ourselves goals such as completing a certain distance of walking and running , or perhaps reaching certain heart rate zones during interval training .
Similarly , in industrial technology , we may set certain goals regarding machine performance . We might stipulate the number of running hours before maintenance is needed , or define parameters around tolerable levels of vibration or temperature ranges . Automation and AI tools are able to track any maintenance issues that are likely to arise .
Sometimes , however , we do not stay within our targets . If we fall behind on our exercise programme , our fitness devices send us alerts , suggesting we carve out more time for a visit to the gym . Alarms and haptic capabilities on the devices keep us continually aware of how we are tracking against our fitness goals .
With IoT-based mining technology , historical data can be fused with real-time sensor data , to warn us when machines are likely to overheat or require new parts . With this ‘ predictive maintenance ’ approach , mine operators can improve their ability to reach organisational targets around uptime , outputs , budgets , and more .
Fitness tracking devices have always dealt with the fact that health care must be personalised . Our bodies are all different and respond differently to different environments . AI in devices create this ability to appreciate ‘ context ’ — knowing that the number of calories one burns can depend on factors such as height , weight , age , gender , and health state .
In an analogous way , algorithms in mines must consider a piece of equipment ’ s age , service history , fault history , productivity records , and so on . Added to this , it needs to factor in environmental factors such as shaft depth , humidity , or temperature . By understanding all these factors combined , we can individualise predictive maintenance recommendations for each piece of equipment .
We have already seen the first inklings of fitness devices bleeding into broader business ecosystems , with medical insurers allowing people to integrate data from their devices into their insurance profiles . They can better tailor one ’ s premiums and use loyalty and rewards to encourage people to improve their health . b
About the author Louise Steenekamp is director of Energy & Natural Resources ( South Africa ) at Wipro Limited .
[ 40 ] MINING MIRROR MARCH 2018