A vessel due at 06:00 now looks likely to arrive at 14:00. Nothing dramatic has happened yet, but for the berth planner the day is already different. If that berth was promised to another ship at 18:00, one of them has to move. Moving a ship means moving crane gangs and shift plans. Export boxes stacked near the berth for the first vessel may now be in the way of the second. Trucks booked for that window will queue at a gate with nowhere to send them.
Planners deal with this all the time, mostly by phone and experience. What they often lack is a quick way to see the consequences of each option before choosing one. That’s a fair job for a port twin.
Three data sources, and what each can tell you
AIS gives position, speed and heading from the ship, plus a destination and ETA entered by the crew. Treat the position as useful and the declared ETA with suspicion, since a person types it and it can go stale. Build your own estimate from track, speed and the history of similar passages, including time spent waiting at anchorage. Report it as a range. Expect gaps where shore receivers lose coverage.
The terminal operating system holds the berth plan, yard positions, planned moves and gate appointments. It’s the backbone of any twin, and it is also a record of what people intended. Containers get moved and not recorded, so check it against reality as you would any other source.
Crane and PLC telemetry tells you which cranes are working, how fast, and where faults keep recurring. Plans often assume a standard crane rate. A rolling figure from each crane is better, and a crane with a repeat fault changes what a berth can really handle tomorrow.
Add tide and weather where they matter. A small delay can push a deep-draught ship past its tidal window and cost far more time than the delay itself.
What the twin actually does
Strip away the 3D view. The useful core is a fast simulation that takes the current state and asks what happens next. If we hold this berth, swap these two ships or add a crane to the next shift, what happens to berth time, yard congestion and gate queues over the next day or two?
Planners compare the options, and they still make the choice. Trust depends on showing the assumptions, so run the model against past days first and check whether it would have predicted what really happened.
Keep it in proportion. If your terminal has one berth and a handful of calls a week, a spreadsheet and a good planner will do fine. If your yard records are unreliable, a simulation will give you wrong answers with great confidence, and fixing the records is the better first project.
One way to improve yard data is to stop relying on manual updates. AeroTrack follows containers, chassis and reach stackers across yards and quays, fully wirelessly, with precision of 2 to 5 metres and a range of up to 7 km per gateway. A twin can then work from live positions.
A first step
Pull six months of vessel calls. For each one, compare the ETA you held 48 hours and 24 hours out with the time the ship actually reached the berth. Then check how many of those misses changed the berth plan. The result shows how big the problem is, and whether predicting arrivals or simulating your response should come first.