Colloquia

Summary

The RevisieLoopwerken or RLW maintains the boogies of the trains of the NS. The bogie is separated into two main parts, the wheel axles and the rest of the bogie.  The wheel axles of the bogie are maintained by a production line in the city of Haarlem. The production line however does not perform as it was designed to do. The goal of this research is to find the limitations of the factory and how the overall throughput can be increased.

There are multiple causes to the performance of the factory. The problems are partly due to the planning of the different axles, the performance of the machines, and the lack of insight into the result of a change in the routing.  There are  17  different types of wheel axles.  There are two main groups each axle can be distinguished to, namely rolling axles and axles with a gearbox. 

The different aspects of the production line were included in a simulation model. In this simulation model, all the different routes for axles were put in with the help of a process flow. Data from the factory itself was used to set production times, the distribution of axles types, and the failures of machines. Workers were also added to perform different tasks. Each worker type has a specific set of tasks the worker can perform. 

The simulation model was validated with data from the factory. With the help of the simulation model, different scenarios were then tested in the factory. Here some changes were made to the ratio of gearbox axles and rolling axles in the factory, a possibility of a night shift, number of workers, new planning rules, and extra machines. 

 All the results of the experiments were tested against the normal situation. The simulation model confirmed the importance of the ratio of gearbox axles and rolling axles. A new planning rule, where no more than 2 axles are put in sequence, was found to have an increase in throughput. Another increase was found by adding one extra worker of a certain type and an even larger increase was found adding an extra decoating machine as well. The increase in throughput was also found by adding extra working hours for workers. 

 It is recommended to update the data used for the distribution of axles regularly. The distribution changes over time and therefore the experiments should be tested again for expected future distributions. In these experiments, other factors might be the limiting factor. The simulation model will be used to test new planning rules and new routes. It will also be used by production engineers to increase the throughput and gather insights into the effect of changes.