Published on 11/28/2016 | Strategy
After reading tons of articles about situation in mining and how its fortunes seem to ebb and flow in somewhat predictable cycles, I believe that the next recovery in mining might not be like any before. We will still continue mining the resources, there is no doubt, but how we will do it will change forever. While becoming smarter, mining will also have to be more attuned to the costs of environmental damage it is inflicting, to the much higher cost of labor, and to the cost of digging in marginal locations far more remote than ever before. The key enabler of changes will come in form of a wave of digitization sweeping across countless industries.
Mining is not the first industry to experience seismic shifts due to technological advancements. The same phenomena happened in manufacturing, even acquiring its own moniker: Industry 4.0. Not a rigid manufacturing methodology, but a coherent response to rapid increases in on-demand computational power, omnipresent connectivity, emergence of analytics and data-driven optimization, new forms of human-process interaction such as augmented-reality systems, and advancement of control systems and devices (IoTs) communicating digital instructions to each other.
Mining has to experience the same shift as manufacturing. There are fewer new mines being built, many existing mines are producing lower ore grades and require longer haul distances from the mine faces. The perilous state of the industry is showing up in decreasing productivity of mining operations globally, something that has not escaped attention of the bankers financing existing mining operations or measuring up new projects. Even if mines are not factories, but a collection of individual projects, each with its own distinct DNA blueprint consisting of location, geology, and logistics, the change is inevitable.
The exploration is getting very mobile thanks to advancements in portability and computing power of lab analytical equipment and extensive usage of autonomous drones. That enables identifying minerals, chemical compositions and physical properties in the field, while feeding analysis results back to the desktop mine simulation and economic asset modelling programs. The field-back office integration feeding in parallel the processes of chemistry analysis and financial analysis, solves the prescient problem of mining: knowing what is in the ground, where is it, how much is it really worth, how much it will cost to get it to the market, and what is the actual market demand for it.
The mining operation is as exciting and will see a wave of changes. The current usage of sensors on the equipment feels still like "20th century". It could be summed up as “report back to base & let the human decide”. Thanks to advancement in robotics, intelligent control devices (IoTs) and equipment-to-equipment communication, blasting, drilling, shoveling, and haulage are changing in many ways.
As the machines will communicate with each other and move into making intelligent machine-to-machine decisions, demand for human operators and human controllers will decrease. The remaining workforce will become far more advanced and the profile of their skills will change from machinist to cyber equipment control specialist. The next step will be even more profound, as machine learning becomes embedded in computing equipment on board of the machine itself, instead of residing in computers at remote computerized control centers.
Moving from sensors to intelligent IoT devices is already enabling real-time flow of information that not only provides humans with better insights early, but also allows automated replanning and rescheduling of activities in the mine based on financial and operational factors. This gets away from heuristic decision making based on past experience and moves to machine self-learning and refining each next decision with richer set of data. That’s where computing power available at low cost comes into play, as more data generated by the machinery and sensors can be iteratively processed faster and with much higher precision than any human can dream of. On the broader scale of pit-to-plant-to-port, the decision making algorithms not only digest in real time the data flowing in, but can also produce new data that becomes valuable in itself for the purpose of self-learning, something already known in the manufacturing industry as self-learning supply chains.
The robotics and machine learning changes the economics of production. At Quintiq, we work with the largest and most efficient steel producers in the world, an industry comparable to mining in its capital intensity. There, the changes to how continuous production lines are being automated, how line control devices interact with each other based on quantitative massive data processed rapidly for continuous planning decisions by computer-based optimization resulted in machine productivity of nearly 100%, a far cry from machine efficiency seen in mining.
In essence, our steel production customers abandoned tools and metrics narrowly focused on overall equipment effectiveness (OEE), that don’t grasp complexity of the whole value chain. They understood, just as Industry 4.0 advocates, that their supply chains are interdependent systems of multiple pieces of fixed and mobile equipment working on material featuring dynamic changes in chemistry. Real-time data and better mathematical optimization engines made possible scheduling and processing decisions that significantly increased utilization of equipment and yields. Applying the same logic to mining processing plants, algorithmic optimization eliminates blind spots in understanding the drivers of yield by looking for hidden relationships between distant variables, something that heuristic optimization cannot accomplish.
Another valuable aspect of Industry 4.0 is that it does not concern itself only with robotizing the production, but also with the environment in which it operates and the waste it produces. That in itself is incredibly valuable to mining, as data-based optimization and algorithmic decision making can lead to far more sophisticated management of mine drainage, acidic sludge holding, and transforming tailings into beneficial products. The costs of doing too little or without much sophistication can be seen in the penalties assessed against companies like Vale, BHP, or Rio Tinto following major environmental impacts that could have been prevented given the sophistication of control equipment, sensors and decision-making algorithms available on the market today.
- Pushing intelligence to IoTs autonomously controlling the machines increases security issues, as hacking into one of the devices gains access to the whole network of interconnected devices. If controls are located in the 3rd party cloud, complexity of fighting the intrusion increases greatly comparing to “on premise” control and optimization centers. Read this post for better understanding of the problem.
- Local and wide area networking needs high degree of reliability and fail over capabilities, something that may be difficult to ensure consistently in remote terrain with unfriendly climate.
- Many high-paying human jobs will inevitably be lost due to advanced automation of processes and decision making, even as the remaining jobs become far better paying.
I manage Asia Pacific division of Quintiq, a constant innovator in applying mathematical optimization to solve real life business planning & execution problems. All my writings draw on real life business experiences with my clients. Asian examples feature big, because I live and work in this region and see its dynamics first hand. If that interests you, please follow me to receive the latest updates.
This article was originally posted on LinkedIn.
