Northern Border Pipeline simulated every pipe move and rented 20% fewer railcars

Northern Border Pipeline needed to move enough pipe to complete a pipeline within a planned nine-month window. The obvious defensive move was to rent extra railcars so construction would never wait. That would protect the schedule, but it would also hide poor coordination in expensive idle capacity.

Deepak Bammi built a GPSS simulation of the pipe logistics system. It was detailed enough to include pipe arriving from abroad, pipe made in domestic factories, train loading, transit and breakdowns, barge legs, unloading, inspection and welding at the construction front.

The simulation showed the same deadline could be met with 20% fewer railcars. The reduced fleet was hired, the pipeline was completed on schedule as predicted, and the project saved $4.5 million in railcar hiring costs.

Simulate the bottlenecks before buying capacity

It replaced safety-stock thinking with a working model of the whole construction supply chain.

It included real disruptions such as rail breakdowns instead of assuming perfect transport.

It converted a schedule-risk problem into a capacity decision that managers could price.

When buffers are expensive, model the flow before renting the buffer. The right answer may be less capacity plus better timing.

The published Interfaces case reports that the model saved $4.5 million, cut required railcars by 20%, and produced insight into how factory operations, loading, transit, unloading and welding had to mesh for the schedule to hold.