As exhaust emission regulations get tighter, and fuel gets more expensive, the need for the marine industry to find more efficient ways of operating vessels is becoming ever more critical. With a large vessel using 150 tonnes or more of fuel a day, and a heavily fouled ship consuming up to 40% more fuel, it’s a significant issue that adds vast costs to vessel owners and operators.

Not just about fuel consumption

The problems of hull fouling stem from the fact that the hull is rougher, increasing drag, but also that can it adds a considerable amount of weight. However, the problems caused by fouling are felt long before larger organisms start adhering and growing (hard macrofouling) – once a so-called biofilm begins developing (microfouling), fuel costs may have already risen by over 10%.

And it isn’t just about fuel consumption – dry docking to clean and coat vessels is expensive and time-consuming both in itself and because in dry dock, the vessel isn’t creating revenue.

Complex and dynamic

Given the level of active research, why is fouling still one of the most significant cost factors in operating a vessel?

One of the reasons is that, even in the 21st century, the interaction between marine organisms and hull coatings is very complex, dynamic and not well understood. It involves thousands of different species that range in size from 10nm to 10cm.

Vessel owners and operators want to understand:

  • How best to operate a vessel as efficiently as possible as fouling occurs and decreases vessel operational efficiency
  • The optimal dry dock schedules, and being able to plan them well in advance to minimise their costs and to integrate them into shipping schedules
  • How vessel cleaning processes, and coatings types, impact the performance of a specific vessel, so that informed decisions can be made

Machine learning

Traditional approaches to counting the costs of fouling have typically amounted to some form of linear regression. But as we know, reality – including fouling – is not linear. In reality, fouling processes are much more dynamic – for example, certain species of macrofouling organisms will typically settle within hours of a vessel returning to the ocean from a dry dock.

Given the complex, dynamic and diverse effects of the factors involved in analysing the impact of fouling on vessel performance, the GreenSteam machine learning approach is particularly successful at providing insight into the areas that owners and operators are most interested in.

Machine learning takes all factors into account

Machine learning can factor in:

  • Vessel type, size, design, and speed
  • Vessel routes and idle periods
  • Sea temperature, geographical location and season

By combining what we know about a vessel, how and where it has been operating, and what its dry dock, cleaning and coatings schedules have been, the machine learning approach can deliver a more accurate fouling analysis than traditional, linear, methodologies.