Machine learning is great for mining, but not a panacea
The advances in artificial intelligence technologies, mainly driven by ML development, certainly offered new efficient tools to solve the problem of ore reserve detection.
In fact, previous approaches based on geostatistics techniques struggled to manage the typical mining datasets, sadly famous for their properties of complexity, sparseness, and impreciseness.
The application of ML algorithms, such as support vector machines (SVM), and the intensive use of neural networks helped mining engineers cope with all this variability, in order to combine different layers of data and identify key relationships among them.
The fact remains that even machine learning-based approaches require quality data to train machines in the best possible way. And that’s not always the case.
Historical data can be messy, inaccurate, or limited. We may have abundant chemical data but little information about magnetics and radiometry. This can be challenging even for the most advanced ML algorithms.
2. AI reduces mining environmental impact
Drilling around usually involves making a big mess. A big and dangerous mess.
The environmental problems regarding mining include the formation of sinkholes, erosion, a reduction of biodiversity, the contamination of soil from leakage of chemicals, and so on.
Mining by its nature is an extremely destructive process, but the introduction of AI technologies is dramatically reducing its impact on nature.
For example, the widespread deployment of sensors and cameras inside and outside the mining sites ensures to monitor excavation and extraction activities, detecting any leaks of waste and harmful materials.
Our “cybernetic eyes” can be integrated into AI and ML systems to automate video surveillance. This allows them to autonomously analyze large volumes of data and spot some anomalies or shifts from the usual patterns they are trained on, such as temperature changes and tremors.
Basically, an excellent tool for avoiding accidents or environmental disasters.
Algorithms optimize mining energy consumption
Another way AI helps protect the environment is by reducing the power consumption of the mining industry.
Ventilation represents the main energy cost in the subsoil mines. You know, miners without air cannot breathe. And people who can’t breathe, die.
Even if some Minneapolis cops seemed to think otherwise.
In this regard, AI-based automatic regulation of ventilation systems is a key factor in saving power, also by using machine learning to forecast energy peaks.
Less decisive but still useful for saving purposes is the automation of the minerals sorting. During mining operations, larger amounts of materials should be removed to find the valuable resources we’re searching for.
Some companies started to use AI-based machines to sort useless material from the cool and shiny stuff, saving fuel and energy during the process.
3. How algorithms can protect miners
Unfortunately, the mining industry is not only a potential danger to the ecosystem.
Although in recent years it has become more fashionable to think about the environment than working conditions, mining accidents involving personnel still occur quite often.
According to government estimates, about 5,000 Chinese miners die in accidents each year, while independent reports speak of over 20,000 victims.
AI is a good friend for miners
Artificial intelligence ensures better and faster decision making: an essential prerequisite for improving the health and safety conditions of miners.
It allows us to collect in real-time a wide range of data from the sensors and cameras mentioned above, detecting anomalous and potentially dangerous events.
These tools can also be paired with wearable sensors, which continuously monitor worker behavior to spot any sign of physical discomfort and send an alert.
4. Drones and automated machines for mining
As we have seen, mines are dangerous places due to accidents, contamination, fires, balrogs, and many other terrible things.
Even zombies, according to the Minecraft video game.