Is Mining about to go Micro - Post 3

In Part 3 we look at some of the technologies that will underpin automation.


If you could have smaller vehicles operating independently and automatically you might be able to make an economic case for this scenario. Of course you'd need to implement all of the technologies you need to operate a large number of autonomous vehicles. I'll cover a few such technologies here and provide some idea of where that technology is today and where it might be heading.

Advanced Analytics

The use of analytical algorithms that assist with making many low level decisions is quite advanced in many industries. Because the mining industry is not as technologically mature as others, there is plenty of opportunity to learn from how, say, the aerospace industry uses these advanced analytical techniques to remove the need for human intervention in many processes. Aircraft can now land themselves on autopilot, a specific and highly complex event that needs microsecond by microsecond adjustment of aircraft and engine power in three dimensions, and responding to highly dynamic changes in the surrounding environment in real time. All of this can happen with a plane load of hundreds of people. As a specific task, they don't come much more complicated, and this level of knowledge is in use today in commercial airliners. Many of the technologies and techniques are available for use in the mining industry today.

Smart sensors

Many of the individual, microsecond decisions can actually be undertaken away from the smart core algorithm by smart sensors operating at the periphery of a complex computer algorithm. For very well understood and constrained processes a smart sensor might ‘decide’ to provide data to a central algorithm only when the process deviates from a pre-established norm. For example, a smart thermometer might only report fluctuations in engine temperature outside of a particular range, or if the temperature changes much faster than expected. This takes pressure off the central processing ability of the system.

Machine Intelligence

Both advanced analytical capabilities and smart sensors lead the way to 'machine intelligence'; the ability of a machine to have some level of understanding of the environment around it and of its role within that environment. At the moment, small mobile machines are being designed with levels of intelligence approximating insects. Indeed, some of these machines are incorporating insect neurons within the machine's IT architecture. These machines learn!

At the moment, these machines are even programmed to behave like insects - seeking areas of warmth, running away from the light etc, which might not immediately seem of use to the mining industry, but that is not the limit of what they might be ultimately be capable of.

Automation

With all of these things being either in existence now, or being actively researched, machine autonomy may not be very far away at all. Indeed Komasu and Caterpillar have already produced haul trucks that can be operated remotely, and are working on being able to operate autonomously. Work at the CRCMining (www.crcmining.com.au ) has some machinery automated for some of their functions. To help cope with the communications lag times the Mars Rovers have limited levels of autonomy to allow them to move around on the service of Mars with minimal intervention by people.

Robotics

Another technology that is critical to automating machinery is robotics. Robotics incorporates all of the other technologies discussed here, but also concerns the form and function of the machines. Robotic science is very highly advanced and very sophisticated. Some of the recent advances look to nature to inspire the development of robots, and it is here that I believe the mining industry can look for the future of the mining process.

When Safety Systems are unsafe

I was looking at a press release by a large mining company that, rightly, expressed frustration and disappointment that they had not been able to meet their target of 0 deaths in the operation. All mining companies I know of are certainly aiming for this target, but very few can do it. This is true even though there are very strict processes to follow to continually improve safety outcomes.

So what gives? I believe that the focus on safety needs to be reassessed. Yes, focus on safety, but look at it as one part of a whole. If a company designs a safety system that does not also help improve production outcomes, then the first time the company puts on the squeeze for higher production, then people will compromise on safety. And they will do it on purpose and with a pretty clear understanding of the possible outcomes, and how likely they are to happen. The chances of this behaviour happening is increased if a persons pay packet is liked to production.

So here is my take - design systems that deliver better safety AND production outcomes.

Reposted from "Intersections" http://miningit.blogspot.com

4000 metres deep and counting.

A couple of weeks ago, my colleague Colin Farrelly and I travelled to Johannesburg in South Africa to participate in a "Technology Innovation Consortium" with AngloGold Ashanti - http://www.aga-tic.com/agatic/.

The idea of this group of invited participants is to redesign the mining process, from the ground down (if you like). We have a marvellous time brainstorming solutions to really hairy problems The fundamental issue facing AngloGold Ashanti (AGA) is that at depths of greater than 4000 metres the working environment and the logistics and energy requirements to operate become very difficult indeed. Col and I went to the bottom of the mine, 4000 metres below the surface and I have to say the trip was an eye-opener. I've been to a few underground mines before, and they pretty much look the same once you are below the surface, but this one is a doosy. Just the thought of all that rock above you makes you take a pause for reflection.

For a start the temperature of the rock down there is 65C (160F), and at 5000 meters, where the company hopes to be able to mine in the future, the rock temperature is 85C (200F). In the presence of up to 100% humidity, people cannot work at these temperatures. Enormous amounts of energy are spent on making the environment livable for the workers, and at greater depths, even more energy will be needed if current methods are continued.

In addition, the logistic challenges at these depths are enormous. Large amounts of water are pumped from the top to the bottom of the mine, and even more is pumped out (groundwater leaks in). Thousands of tonnes of ore are lifted out of the mine every day. On my visit it took me nearly 2 hours to get to the working face, and the same to return. Its like that for every miner as well so of their 9 hour shift, 4 hours is spend travelling underground.

AGA have decided that they will not be able to mine at greater depth safely (only a week after we were there there was a fatality in the mine) nor will they be able to use current methods economically. This consortium is a real expression of their desire to change with new ways of mining, automation, innovative use of energy etc. Its pretty exciting and all of this innovation will fundamentally depend on robust, reliable, and safe IT and communications systems.

Have a look at the website to get more of an insight into where the mining industry is going in the future.

Mining underground in Mponeng Mine, South Africa

Automation in Mining

Most mining companies are about to embark on the journey towards ‘automation’. Companies like Rio Tinto have made a good start, but they all have a long way to go. Automation is much better progressed in many other industries, especially where automation of fixed plant is where they have had to focus: manufacturing and other process industries, Oil and Gas and other chemical type industries, and rail / transport systems. Mostly these are industries in which the processes are constrained by the engineering. The challenge for mining is very different, very few mining operations are constrained by the engineering, indeed they are geographically unconstrained (at least within the boundaries of the ore body).

One way to look at automation is through the lens of the future objectives – what is the mining process to look like? Currently the major constraint is that the mining process is essentially a batch process; drill, then blast, then collect the ore, then load it into the transport system. Current automation efforts are really about pushing the continuous process as close to the mine face as possible. What we see people like AngloGold Ashanti doing is investigating what technologies are available to make the actual mining process underground continuous. They seek to get out of the drill and blast paradigm. This is the future of automation.

In the meantime, there are things that can be done to extend ‘automation’ into other parts of the mining process, but for every company, that will mean something different, so each needs to define their automation objectives and a high level view of how to get there. These matters need to be considered (in order) as follows.

Automation Objectives. Mining companies obviously need to understand what problems they are trying to solve – nobody else can really do this for them, although the AGA Technology Innovation Consortium (AGA-TIC) has taken an ‘open innovation’ approach. (www.aga-tic.com)

What is Future Vision. Every journey has a first step. Understanding the future state is informed by the objectives and needs to be developed without constraints in the first instance, that is don’t worry about where you are today. Big innovation will probably replace the current state anyway.

Technology. Future technology is perhaps the hardest thing to quantify – current trends only loosely inform where technology will be in even the medium term. The trick is to plan for the no-brainers and build agility into the technology plan. Any automation story will however be underpinned with better and more pervasive communications, smarter ways of collecting and analysing innovation, and robotic technologies that will drive new machines.

Culture. Culture will be the most difficult thing to change, and will take a long time. Companies can’t afford to lose the knowledge in the current workforce, and anyway new technologies are driven by the young, and they will take quite a while before they are impacting how you do your mining business.

What is Current State. Once the end state is defined, you need to know how ready your current environment is for the changes that will be necessary. Knowing where change is needed and where it can be ignored is all part of doing an Automation Readiness Evaluation. Besides, there will be some things that can be done early that will improve the business and help to finance longer term initiatives.

Technology Innovation Plan. How you go about innovating is the core of the implementation problem – will it be open (like AGA) or closed (like Rio Tinto). Both of these methods require investment, but open innovation will probably get a faster result.

Technology Vision. Prototyping automated technologies, Innovation in iterations...

Systems Integration. Making sure that the systems you design can act in awareness of each other is a key to machine automation. Machines will need to negotiate with each other in the mining environment because they will have freedom of movement. For instance, to truly automate a haulage fleet, each machine will need to know its own place in space and system, as well as every other machine. This means a common language, decision rules, etc.

Knowledge Management. Automation is really a way to embed intelligence that currently resides in people’s brains into the systems and machines of the future mine. There is a huge cultural change issue in this.

Further Observations

Many industries have been on this journey before – mining companies starting out should make themselves informed on the lessons they have learned – petroleum, aerospace, military in particular.

As a concept of operations, automation in any context will ultimately be made possible by how well you can collect, transport, and analyse data automatically.   We currently know a great amount about this, and recent advances in intelligent sensors, mobile communications and advanced analytics are making the issue more addressable – complicated to be sure, but the technologies are becoming ever more capable.