Q&A: GreenSteam talks how machine learning can reduce sulphur emissions

Varsha Saraogi 21 November 2019 (Last Updated November 21st, 2019 15:16)

Mitigating emissions to adhere to the sulphur cap regulation could be catalysed thanks to a new machine learning platform developed by GreenSteam, says COO Simon Whitford. We hear more about how the technology can help cut fuel waste and emissions.

Q&A: GreenSteam talks how machine learning can reduce sulphur emissions
Can machine learning solve shipping woes? Credit: Pixabay.

The shipping industry is fraught with challenges. With the International Maritime Organization’s (IMO) 2020 regulation on sulphur emissions just months away from being implemented, shipowners are under pressure to ensure they comply with the new cap.

While many are launching a slew of solutions such as switching to sustainable fuels, slow steaming and installing scrubbers, vessel performance optimisation firm GreenSteam is bidding to help the industry through its machine learning platform.

The platform uses technology to collect and analyse data that is used to predict a vessel’s fuel consumption in the event of climatic changes. Through the data collected, it then estimates the speed at which a ship should sail to reduce fuel waste.

GreenSteam’s COO Simon Whitford claims the technology can be crucial in helping ship operators mitigate emissions in ways that are more efficient, cost-effective and have a lesser impact on the supply chain and voyage duration.

As he puts it, while using low-sulphur fuel is one of the objectives, a “data-led approach to maintain voyage optimisation” is essential going forward.

Varsha Saraogi (VS): What are the main challenges facing the industry when it comes to increasing sustainability?

Simon Whitford (SW): The world shipping fleet is still operating in a post-2008 crisis market and over-capacity still flattens earnings. This obliges shipping companies to focus scarce funds on mandatory compliance projects like BWT [Ballast Water Management], Exhaust Gas Scrubbers etc., with fewer opportunities to fund other sustainability improvement projects, despite the excellent returns often available. This opens the door for shipping companies to engage with GreenSteam’s zero-capex machine learning platform.

VS: What support can the IMO provide to make shipping more sustainable?

SW: As much assistance as possible to help pioneering GHG reduction technologies to develop. Other ways of improving efficiency would be to sponsor trials, audit and endorse results – all these things will help and encourage shipowners to invest even more in their sustainability agenda.

VS: Is reducing ship speed a good way to cut emissions?

SW: Optimisation of speed offers a better decarbonisation strategy than a simple speed reduction. If all the world’s cargo vessels suddenly slowed down and simply took longer to arrive – we would need more ships, and we would have to ask the world’s factories to increase production to fill our new slow supply chain. This might just result in an increase in emissions overall.

GreenSteam’s Speed Optimisation algorithm employs a machine learning platform and runs hundreds of thousands of simulations along a particular route which considers how the vessel responds to the latest forecasts of weather and sea state to recommend a speed profile across the voyage that delivers the lowest fuel consumption specific to the ship, whilst respecting the vessel’s original ETA [Estimated Time of Arrival]. So, optimising speed to still meet the required delivery date, rather than simply reducing speed, will deliver a greater reduction in GHG emissions.

VS: How can machine learning and data analysis help improve sustainability?

SW: Fuel consumption is the input and emissions are the output. Machine learning is the only means currently available to join all the dots to accurately capture how each of the many factors which affect a vessel’s fuel consumption interact over time in very complex ways.

By learning this for every vessel, machine learning quantifies the drivers of fuel wastage. Accurately dividing a vessel’s fuel wastage into each optimisation area such as hull fouling, trim, speed, and weather allows appropriate action to be taken, and fuel wastage savings can be validated.

Every shipowner today has access to high leverage, capex-free machine learning technology to significantly reduce fuel wastage and emissions.

VS: How is GreenSteam helping shipping companies cut emissions?

SW: GreenSteam’s machine learning platform learns how the fuel consumption of a vessel is impacted by all the operational modes and environmental influences affecting the ship. Most of these factors are inter-related and change constantly. This is a highly complex system which is specific to each vessel – one where only machine learning has the capacity and accuracy to measure and then predict each area of fuel wastage. Once the individual components of fuel wastage have been identified, targeted actions can be taken to reduce fuel wastage and cut GHG emissions. It took GreenSteam’s team of oceanographers, naval architects, data scientists and programmers over a decade to build our machine learning platform.

[For example,] Danish logistics company DFDS deployed GreenSteam on their RoRo vessels.

In terms of the technology – GreenSteam’s services were prototyped and further developed insofar the vessel is self-sufficient with our machine algorithm. The platform gathered vessel and environmental data which included draft, power, and speed in real-time rather than from subscription metocean services. This was essential to allow the ship to optimise trim dynamically.

Originally we built a screen interface on the bridge, but for several vessels, this has evolved into a tablet, which can now be operated remotely.

In terms of IoT equipment on board, GreenSteam installed a bridge wing radar for draft and point sea state, inclinometers for trim, and accelerometers for wave state on the vessel. Combined with our machine learning algorithms, this translated into fuel savings and a reduction in emissions.

VS: With more companies using machine learning, how will the process of cutting emissions change in the future?

SW: The challenge is this – the shipping market has a large installed base of legacy technology. Retrofit exhaust gas scrubbers fitted on around 4,000 vessels will dramatically reduce sulphur emissions.

For new builds, I am sure innovative design will deliver dramatic reductions in their carbon footprint. For all vessels, both new or existing, innovative hardware technology and no-capex solutions like GreenSteam’s machine learning platform will continue to offer the shipping industry GHG emissions options.