Detecting vulnerabilities with artificial intelligence

HHLA will check the state of their container gantry cranes with the help of artificial intelligence and photo-capable drones.

When they fail, it gets expensive. The HHLA container gantry cranes can lift up to four containers at the same time at the Port of Hamburg, either onto a ship or down from it. They call this process “löschen und laden” (loading and discharging) at the port. The cranes move tens of thousands of steel boxes in a short time.

To avoid expensive breaks, the enormous, multi-million dollar pieces of metal have to be regularly checked for wear and weaknesses. After all, the containers – which can weigh up to 40 tonnes – place a heavy load on wire ropes and threads, and their steel is exposed to salty water and fierce winds around the clock.

Expensive specialists, professional industrial climbers, regularly climb through the supporting structures of the container gantry cranes with rope, hook and helmet. The experts check every single weakness by hand. This takes time. And the bridges, which are growing ever larger, require more and more checks and inspections.

To simplify these inspections, HHLA will increasingly rely on artificial intelligence in the future.

Troubleshooting with the smartphone principle

The HHLA engineers and technicians want to track down potential weaknesses with the help of artificial intelligence. In principle, this digital troubleshooting works like facial recognition on a smartphone. It learns to distinguish the unique, personal characteristics of the owner from others.  If these deviate from the stored settings, the smartphone denies access.

In the inspection, a drone camera is the smartphone and the giant steel structure is the face. Instead of an unrecognised face, any unplanned wear and tear will trigger an alert that the container gantry crane needs to be inspected.

Technological steps at the container terminal

Weld seams and structural points of regularly used construction components are burdened the most by heavy loads at the port terminals. This is why industrial climbers pay special attention to these points. They are often hard to reach, and a thorough inspection often takes the experts up to two days.

Artificial intelligence can recognise more quickly when rust sets in or if a weld seam is threatening to break. To do this, the drones fly autonomously over the crane and save images of all the important areas. The current images are compared to the target status and the artificial intelligence analyses defects and signs of ageing.

Advantages of artificial intelligence

Advantages of AI: it doesn’t make mistakes. It doesn’t get tired. It is never careless. It is much cheaper. Another advantage: the quality of images taken by the drones is better.

This makes HHLA more independent in choosing inspection intervals. Industrial climbers have to be available – drones can be used at any time.

People will still be necessary

All of this doesn’t make people superfluous at the terminal. If the software finds a suspicious spot, HHLA employees look at the notification on the computer. If the suspicion is substantiated, industrial climbers clarify the situation on-site. If necessary and possible, they repair the defect immediately.

So far, so theoretical. In the first three months of 2021, the system will be implemented at the port terminals. However, the drones are not yet legally allowed to fly completely autonomously.

Click here for more about research projects on port digitalisation.