Developing a locomotive simulation system
Railroad locomotive departments need locomotive planning systems to make strategic decisions related to fleet sizing, fleet mix, shop locations, capacities and capabilities. We developed a simulation system to meet this need for CSX Transportation.
With the variability inherent in train operations and locomotive operations, railroads needed to determine the optimal locomotive fleet size and mix to achieve a specific on-time train departure performance.
Large freight railroads in the United States own several thousand locomotives to power the thousands of trains they operate every day. When trains run late or are cancelled, locomotives become unavailable to run other trains. Locomotives also visit shops for fueling, scheduled maintenance and repairs. With the variability inherent in train operations (with late or cancelled trains) and locomotive operations (with late locomotives or breakdown of locomotives), railroads needed to determine the optimal locomotive fleet size and mix to achieve a specific on-time train departure performance. They also needed to identify ideal shop locations, capacities and capabilities, as well as the impact of different operating practices on locomotive productivity and utilization. To answer these questions, the planners required a simulation system that realistically modeled the movement of locomotives in the train network. CSX Transportation hired Optym to build this simulation system.
We developed a discrete event simulation system to model the randomness and movement of locomotives in the train network. This detailed simulation system modeled train delays, scheduled and breakdown locomotive maintenance and locomotive shop capacities and capabilities. After completing the simulation, we generated a variety of reports related to locomotive utilization and productivity and shop capacity utilization. We packaged the simulation engine within a decision support system, which enabled the user to change input data, run simulations and analyze simulation outputs through reports, graphs and charts. Our new platform also created multiple scenarios for storing and displaying results of different studies. We validated the simulation system by comparing the network performance statistics that it obtained with those observed in the real world. We found both to be sufficiently close.
- Assess the impact of increasing locomotive fleet size on on-time train performance
- Estimate the number of locomotives needed to meet future train requirements
- Analyze the impact of locomotive shop locations, capabilities and capacities on locomotive downtime.
Our new system has provided company management with objective and verifiable insights into locomotive productivity and efficiency.