As one of the largest steel manufacturers and suppliers in the world, Tata Steel faced the need to optimize the internal logistics of one of its steel manufacturing units. The company saw potential for increasing the unit’s overall throughput using AnyLogic as production optimization software.
The steel melting team found that the cranes and ladles of the unit were unequally distributed throughout the production process. This, in turn, had a negative impact on throughput. Therefore, using production optimization software, the engineers tested if changing the existing handling rules of the internal logistics system would improve the situation.
How does the system work?
The production process starts with tanks containing hot metal arriving at the plant. The metal is then poured into ladles and they, in turn, are placed on transfer cars. A crane picks a ladle with metal and brings it to the station, where the metal is purified. When the process is complete, the transfer car moves the metal to a point where a different crane can take it to an LD converter. The crane with the empty ladle then comes back to the tanks.
Until this point, most of the crane handling had been done manually and independently. So, the engineers wanted to find a simple way to operate cranes and optimize the overall production process by reducing the need for human decision making.
The reasons the team turned to simulation modeling were:
- The complexity of the interdependencies between the system elements
- The number of highly random variables (e.g. the processing time of LD converters and desulphurization stations, and the downtime of cranes and transfer cars, etc.).
Using AnyLogic as production optimization software, the team built a model reflecting every step of the process: from the moment when hot metal gets unloaded from tanks into ladles to the charging of the LD converters. They created three different what-if scenarios to reduce waiting time for the converters and modify the crane operation logic. They also designed 270 experiments in total to see how the quantity of heats (batches of steel) produced per day, crane utilization rate, and waiting time of LD converters would vary from scenario to scenario.
The results showed that one of the scenarios had a significant advantage in heats per day and converter waiting time. This meant that implementing it could save the company several million dollars annually. The team was also going to further the production optimization by integrating artificial intelligence into the model.
For the project details, please, read the case study.