The BNSF Railway Company is one of the largest freight railroad networks in the United States. The company has 32,500 miles of track in 28 states and three Canadian provinces, and more than 8,000 locomotives.
BNSF connects freight shippers and consumers in the global marketplace by delivering different types of goods, including agricultural products, coal, industrial products, and intermodal freight. The latter accounts for roughly half of BNSF's total volume, and this share continues to grow. The company’s intermodal hub network is made up of over 26 facilities in the USA, including the major ones, such as Corwith in Chicago, Hobart in Southern California, and Alliance in Fort Worth, Texas.
Expanding the company’s existing facilities, and developing new facilities in intermodal segments, would require making large capital investments. To help evaluate various options of capacity expansion in major facilities, and make data-informed decisions, BNSF developed container terminal simulation models of the Corwith and Hobart facilities.
The team chose AnyLogic simulation modeling software because of its flexibility and Java-based architecture, which helped employees, unfamiliar with simulation modeling but familiar with Java, learn how to work with the program.
Case #1: Corwith Intermodal Terminal Simulation
Problem
For the Corwith intermodal facility, the main goal was to increase the annual lift capacity to keep up with the growing demand. The team knew that new tracking cranes were needed, but at the same time, they weren’t sure if new storage should also be added. The situation was also complicated, as the facility was landlocked by businesses and neighborhoods, so the company couldn’t expand the footprint. The team had to perform one of the following:
- Densify production tracks by adding production tracks in between the existing ones, and then putting taller cranes above smaller ones.
- Convert some of the other areas in the facility into production tracks.
Solution and Results
Another way to expand the capacity was to add a standard rubber tire gantry crane. These cranes are nimble and can move freely through a container yard. However, applying this type of crane does not allow for creating additional tracks. With widespan cranes, a type which allows for more tracks, the process would not be as quick and efficient.
The third type of crane was a cantilevered rubber tire gantry (CRTG) — a hybrid between the two previous types. It is not as large as the widespan crane, it can span multiple tracks, and it has rubber tires, which enables movement between sets of trucks. To see how this crane would influence the terminal operations, the team simulated its functioning in the model.
The container terminal model shows the operational process of the facility. First, the train arrives and then is shoved into the production tracks. Next, a crane unloads a box and then puts it on a truck, which transports the box to another track. After that, the CRTG crane arrives and loads the box to another railcar.
Within the model, engineers tested how the new CRTG cranes operated. To analyze the efficiency of the cranes, the developers included in the model various metrics, such as truck utilization, inventory, inbound and outbound performance, delay time, and others.
The results of the analysis showed that the inclusion of the CRTG crane in the terminal operations would provide the most efficient combination of additional track footage, operational speed, and flexibility.
Case #2: Hobart Rail Yard Simulation — Southern California
Problem
Hobart is the company’s largest facility, processing over one million units per year. When expanding the facility’s capacity, the team of engineers faced the following challenges:
- Landlocked facility was surrounded by commercial enterprises.
- The facility had a large number of arrivals and departures per day.
- Truck traffic at the facility was dense, which is why adding more truck volume could make the situation even worse.
To determine possible bottlenecks and evaluate strategies for their mitigation, the company decided to perform container yard planning and optimization using simulation modeling.
Solution and Results
To find out how to expand the facility more efficiently, the team simulated the operations of the widespan crane and a new type of diagonal parking underneath the arm of the crane. Simulation showed that this combination would work well for this specific facility, in terms of space and the additional flexibility that diagonal parking could provide.
To simulate traffic congestion in the facility, developers implemented the AnyLogic density map, which helped the team understand where the pinch points in the facility were, and find ways to manage them.
The analysis of the model showed how to significantly expand the Hobalt facility capacity and avoid bottlenecks. This necessitated a manifold approach:
- Capital expansion — the company needed to add some tracks and cranes of several types.
- Operational improvements — it should improve the facility operational efficiency through various optimization techniques. One of those was to route trucks through the facility based on the real-time congestion conditions.
- Customer behavior — the team decided to smooth the volume of the facility over the day and week to add capacity without spending capital dollars.
The company is now looking for the automated technology to expand the capacity of their Alliance Intermodal Facility in Fort Worth, which grows 20% each year. Using AnyLogic simulation modeling, the Research & Analytics team has started to work through the routing logic and possible deadlocks, as well as the operations of facility equipment. They will later include this in a full-scale model of their Alliance facility.
Project presentation by Mike Prince, manager of operations research at BNSF
