Modular Mining is a subsidiary of Komatsu and forms part of the Komatsu Mining Technology Solutions (MTS) Team. They use data and innovation to optimize the mining value chain in real time.
In a mine, haul trucks move material from a source to a sink. A source, such as a shovel, fills these haul trucks, which then move the material to the sinks, which are dumps or crushers.
In more detail, the trucks are filled and then travel, stopping at intersections or hazards until they reach their destination. Here, they need to queue, if necessary, then move into position to unload, and finally deposit their material.
Then they can either do some non-productive activities (maintenance, shift change, fueling, etc.) if needed, or they can go right back to the shovels, load up, and start the process again.
Komatsu’s MTS Team wanted to investigate this process to further leverage optimization in the load and haul sequence of the mining operations. They teamed up with SimWell to make a generic mining simulation model for this, as they did not have experience working with AnyLogic.
The load and haul optimization solution that was created works by importing two different files into the AnyLogic model. The model interface shows the layout of the mine with all the connecting road networks. The first scenario file contains details such as where the trucks, shovels, roads, dumps, and crushers are, as well as the materials in that mine.
This is a generic mining simulation model, so the company can use scenarios from different mines in the same model.
The second file has information specific to the simulation, including break and maintenance schedules, timing parameters, and the mean time between failures.
After inputting these into the model, it runs and gives results in the form of KPIs and a detailed log output where queries can be made to get more results. KPIs include total tonnage moved, type of material moved, actual elements of the mine, and so on.
In addition, the company ran several what-if scenarios. One involved changing the intersection management rules. As most intersections at mines are four-way stops, there is an opportunity to optimize this by picking the better truck to go first when more than one approaches this stop.
The model can be run at full speed, and the results are displayed in a few seconds. This is useful when running a number of simulations with different configurations.
As this is a generic mining simulation model, the developers needed to design some rules to ensure that it could be used for load and haul optimization for different mines, as each mine has unique characteristics.
After a truck has unloaded, it is sent to the shovel with the fewest number of trucks waiting to be loaded.
There are many different grades of materials in a mine, but they have been simplified into high grade, low grade, and waste. The high-grade material is sent to the crusher, the low-grade material to the stockpile for processing later, and waste, which is everything else, is taken to the dump.
After a truck has been unloaded, the simulation model will check if any non-productive activities, such as scheduled maintenance, need to be done. There can also be non-scheduled events, such as a breakdown, where the truck will need to stop its activities and go to be fixed.
Extension 1: battery electric trucks for mines
The generic mining simulation model worked well and gave good results, so the developers decided to extend the model further by adding battery electric (BE) trucks for mines.
Trucks are one of the biggest sources of carbon emissions, and one of the options to reduce this pollution is to replace the internal combustion engine (ICE) trucks with battery electric trucks for mines. These are also more energy-efficient and simpler to operate, meaning less can go wrong.
However, they cannot run as long as the ICE trucks before they need to charge. One way to overcome that is by using trolleys, which are energized wires hanging above the road that the battery electric truck can use to charge while traveling.
Unfortunately, not all trucks will be able to charge that way, so a combination of trolleys and charging stations operate.
The developers wanted to investigate the impact on the material moved and other parameters of switching some ICE trucks to battery electric trucks for mines.
This meant additions to the model, such as truck batteries, charging locations, energy use per road segment, etc.
One important addition was a punishment for depleting the battery. In this case, the truck would go to the charging point, charge completely, and then sit there for the rest of the shift.
Initial results showed that swapping the ICE trucks for BE trucks has had a positive impact and that there is room for further optimization.
Extension 2: in-house simulator
The Komatsu Mining Technology Solutions Team has its own internal simulator, but it doesn’t have a UI component. It also wasn’t possible to validate or debug this simulator, so they needed to find a way to do that.
They decided to combine their simple internal simulator with the previously created AnyLogic model. They used a scenario file in their internal simulator, which outputs events. These events can then be implemented in the AnyLogic model and played back. The original scenario file is also imported to enable it to run correctly. Importantly, the model is only showing the events and not making any decisions. All movements and decisions were made by the internal simulator and shown in the AnyLogic model.
Using the AnyLogic model in this way helped their analysts find strange behavior, such as a truck that doesn’t move for the whole simulation or shovels that don’t have any trucks coming to them.
Although no KPIs are available, the team can use this extension to enable debugging of the internal simulator and see what exactly is happening visually.
The first generic mining simulation model was used to investigate load and haul optimization opportunities, and this proved successful. Because of this and the flexibility of AnyLogic to keep building on the tool, they expanded the model to investigate battery electric trucks for mines to identify opportunities for optimization there as well. Finally, they expanded the model again to connect to their internal simulator. This was used essentially to debug the simulator and make visual observations.
In the future, the Komatsu Mining Technology Solutions Team wants to add more complete routing logic for the battery electric trucks for mines, as they have significantly different movements from the ICE ones.
The case study was presented by Kyle Everly, of Modular Mining, a subsidiary of Komatsu, at the AnyLogic Conference 2022.
The slides are available as a PDF.