System can predict container breakdowns

DENMARK: Researchers in Denmark claim to have developed a system that can identify problems with a refrigerated container before they break down.

The transportation of perishable goods around the world requires refrigeration containers that can maintain the same low temperature throughout an entire trip. A broken refrigeration system often results in several tons of rotten fruit, spoiled meat or – even worse – an entire batch of life-saving medicines that are no longer usable.

Although all modern refrigeration containers are equipped with a SIM card and a system that can sound the alarm if an error occurs, it is often impossible to repair a refrigeration container that is already on its way on a ship. So a defective container almost always means damaged cargo.

A system where the containers themselves can predict whether they will need to be repaired – even before they break down – has been developed by Rasmus Lundgaard Christensen from Aalborg University, in conjunction with Danish controls company Lodam, a Bitzer group company.

“All newer containers already have a lot of monitoring equipment keeping an eye on the different components of the cooling system. We’re using this data in a new way to predict when there is a high risk of a breakdown,” explained Rasmus Lundgaard Christensen.

Christensen analysed data from the refrigeration system in containers that have broken down and found a number of factors that can indicate whether the refrigeration system is getting worn down or is in need of new parts.

“For example, there may be a compressor that is starting to compress liquid instead of gas. The system can register something like that immediately,” he said. “We know that if a compressor is doing that, it’s only a matter of time before it breaks down. When we have that knowledge, we can repair the container the next time it’s in port, before it breaks down, and thus avoid valuable cargo being destroyed.”

Expensive testing

Today, all refrigeration containers are put through a pre-trip inspection (PTI) between each journey or if they have been through a repair.

This is a thorough inspection that checks that the system is working, but the process is both time consuming and costly. PTI testing is estimated to cost the shipping industry over €400m annually.

The developers of the new system maintain that if the containers were equipped with an intelligent system that could predict errors this test would in most cases be superfluous.

“If the container itself can predict which components need to be replaced and indicate this in advance, there’s no need to spend time and money testing the entire system over and over again. It’s a bottleneck you can skip,” Christensen insists.

Skepticism

Christensen’s main challenges in designing the new system have not been the data collection or mathematical algorithms to predict a breakdown, but rather the skepticism of his colleagues in the container industry.

“It’s no secret that there is a certain kind of conservatism when it comes to containers. The design of the refrigeration system in containers has virtually not changed in the past 50 years, and now we come along with a system that almost gives the containers artificial intelligence – or at least an advanced form of machine learning. There are a lot of people in the industry who just need a little time to get used to that.”

However, although smart, Christensen admits the systems will not be able to predict and prevent all breakdowns.

“We work with mechanical parts, pumps, gases and fluids that are flowing. It’s not predictable in the same way that a digital system is, but we’ll probably be able to predict 95% of all breakdowns,” he estimates.