Fire sensors are not enough. With the help of big data, for example information from machinery, cargo and cars, fire safety on ships can be improved. A pre-study from Lighthouse and the Swedish Transport Administration has taken a closer look at how this should be done.

At the end of August, a large evacuation operation had to be started on the passenger ferry Stena Scandica, which was running without power just north of Gotland. According to the Swedish Maritime Administration, the cause of the accident was a fire in a refrigerated container on the car deck which, despite being discovered early and extinguished quickly, "generated a so-called blackout, a power cut on board". After a few hours, the ship started one of its engines and the evacuation could be stopped.

“There are quite a few examples of fires on board that start in cooling units on car decks. These can be of very varying quality. Gas stoves in mobile homes and caravans are also problematic. So it's not just electric cars that you need to worry about, but there are many possible sources of fire on board ships”, says Mikael Hägg at RISE, who led the work on the feasibility study AI-based Fire safety system using Big Data, which was done within the framework of the Swedish Transport Administration's industry program Hållbar sjöfart operated by Lighthouse.

The earlier a fire is discovered, the greater the chance that it will end well, as in the case outside Gotland. The problem is that many fires can be difficult to detect before it is too late. So what more can be done to improve today's fire safety systems? This is where the vision of an integrated information system comes in.

“We asked ourselves the question of what information is available on a ship that can partly help detect a fire, and partly provide decision support for fighting it. And that information does not necessarily come only from fire sensors, but is much broader. It could come from other ship systems, such as ventilation and doors, but also from connected cars and cargo.”

How fires on container ships and RoRo ships are detected seems to differ. On RoRo vessels, accident investigations of several cases show that they are detected quite early by humans or detectors/alarms, while all fires (five cases) on container vessels were detected later and then always by humans.

“They are discovered when the crew goes on watch rounds and so on. One problem with containers is that you know very little about the cargo. You know if they contain dangerous goods, but not much more. We were also able to establish that weather conditions are often a contributing factor to fires sooner or later occurring on container ships”, says Mikael Hägg.

Therefore, the researchers believe that a concept with smart and connected cargo will be an important part of the fire safety system of the future. It's a sensitive area, as cargo owners don't necessarily want to reveal what's in their containers.

“If information is spread that this and that container contains lots of expensive computers or televisions, then of course there is a risk of theft. There are security aspects, commercial aspects and legal aspects to this that need to be resolved.”

Another challenge is to integrate all the information without alarms everywhere all the time.

“It is called alarm management. The operator must understand when things are really going off and get a good overview of what is happening. With big data, there are great opportunities to create a decision support system and eventually perhaps a fully automated fire safety system”, says Mikael Hägg.

Footnote. The pre-study AI-based Fire safety system using Big Data has been written by Mikael Hägg, RISE, David Schmidt RISE, Johan Gregersson, Chalmers and Kenth Sandell, Consilium Safety Group

 Read the report