Fleet management systems have been around in mining for a long time and they are complex expensive beasts that require a considerable amount of effort to install. This has always been an accepted fact but then came along Uber and ride hailing and the world of mobility changed forever.
Before Uber we had taxis. They were a particular type of vehicle and they had a dispatching system in them with taxi ranks located at airports, shops and in the city. Despite all this, taxis were expensive, never on time unless you were at a specific taxi rank and just try and book one for an airport run. Now, you flick up an app on your phone just before you want to leave, they arrive you go to your destination and you get out. I no longer hire cars anymore when I travel as it is so convenient and cheap to use ride hailing. How successful is this? Uber is worth more then any mining company in the world.
Project 412 is now developing Project Rock and looking for interested parties to join us on this journey to rediscover the best way to move rocks in mining.
When I looked at the future of fleet management and short interval control awhile back, I mapped out what I thought would be valuable from sensors mounted on existing heavy equipment, such as truck and excavators/shovels in the open pit environment. I even scanned the industry to see what companies where doing. The missing step in all this was the lack of edge computing, which is important for both data fusion (fusing data from more than one source) and providing real time data outputs (usable data that can be digested by existing software packages). Whilst machine automation may bring this in time there is no reason the mining industry couldn’t have sensors and edge computing on machines right now. We currently collect discrete pieces of data from discrete sensors and then fuse the data into various pieces of software back in the office. It is people intensive, time consuming and delays the use of valuable data. Bolting on sensors in the field with edge computing that feeds directly into mining software via edge computing in real time could be the next big step in automation productivity for mining.
Data block diagram for edge computing. In this diagram I put together some time ago, I looked at how sensors on mining equipment equipment might work with each other to collect the data. The reasons for mounting sensors on key mining equipment are twofold, one is that key pieces of mining equipment are usually at the interest points in the mine where you want to collect data and secondly, high quality sensors negate the need for a physical visit by geologist, mining engineer or the like, pushing towards no entry mining. The on-board edge computers would allow for remote access to the sensors and data that would act independently of the piece of equipment’s primary duty. In this model not all the data is sent back to the control room or office, it is stored on an AWS snowball but can be accessed as required. The Snowball is a remote cloud storage device that you would empty every few days during a shift change so as not to clog up the available bandwidth on site.
Global Data Plan
In this diagram I wanted to highlight how easy it could be to create a system, just using an API platform (think of your phone and its apps) that would remove the need for complicated, bespoke FMS and costly installs. Such a system would also ensure it would be cost effective and cheap to swap one system out with another. I did do some API mapping a couple years back and I am still pretty sure you could manage the whole material movement process in mining with less then a hundred API’s. For example, you could get a vehicle health system from RCT or Dingo, do some geology inputs with Seequent, feed in some data from Deswik into your ‘rock hailing’ or FMS, modify Hexagon MineSight reconciliation and update Surpac – just using API’s. If you don’t like one piece of software, change it out with another without spending a fortune. Basically anything you look at and then proceed to take an action on, you can create an API for. Data change = Action and Actions = data change