Photo credit: Pexels

AI is an ideal tool to extract answers to questions that the maritime industry has historically lacked the time to fully consider, writes ECOsubsea Chief Sustainability Officer, Abigail Robinson.

Shipping is sitting on a potential goldmine in the form of unanswered questions about the best way to optimise operations that most people and organisations don’t have the time to answer even if they have the data to do so. Unless they are mission critical, these questions have historically been deemed as not needing to be answered since the focus is on the day to day rather than an amorphous future. 

But AI holds the keys to being able to clean and sift through data at unprecedented speed, making it possible to answer questions about routing, fuel consumption, the workforce, machinery and more – in a matter of minutes in most cases.

This technology could be harnessed to identify life, planet, or money saving opportunities nestled amongst that mountain of raw data. This is not a theoretical concept but something that I’m seeing happen in real time across our industry – and that I’m working towards myself.

Opportunity and action

A few months ago, I was struck by the thought: “What if all of these unanswered questions are quietly draining millions of dollars every year via a ship’s fuel tank and racking up a hefty GHG bill? Worse yet, what if the answers are sitting on that giant spreadsheet that someone told themselves that they would play around with when they have a spare minute?” 

The idea that the key to solving a million-dollar problem was in our hands all along is usually enough to make anyone pause, but until access to this technology, this thought usually led to a person putting that data analysis hunch on a list of things they would “get to when they have a spare minute”. In most cases that hunch would be forgotten, but even when it was acted on, the time that it took to act on that hunch was as a value leak. 

In some cases, the reason that this question would remain unanswered is the belief that it would be time and energy intensive to find the data and perform the analysis. We need to switch out of this mindset to really take advantage of AI.

Application in the real world

Let me give you an example from my first hand experience. In the slimy and crusty world of biofouling, this kind of heavy-hitting impact potential is being unleashed by data that is usually disposed of or ignored; the biofouling waste collected from a ship’s hull via in-water cleaning. The volume of biofouling waste, securely captured to spare the local port from pollution and invasive species, gives a reliable figure. This in turn opens the door for predictability around fuel, pollution, and invasive species impacts. 

By linking together the weight of biofouling removed during cleans and fuel performance data analysis pre- and post-clean, a new level of hull management strategy becomes available. This value provides concrete insights such as fouling density and growth rates, which in turn are linked to fuel impacts. This type of insight was historically inaccessible due to biofouling assessment being based on visual scoring, which is a highly subjective method.

Thanks to a data sharing collaboration with a close customer, my team and I have been able to predict fuel performance impacts based on a standardised biofouling density score obtained following a clean. In fact, we find this method to be accurate within 0.5% compared to customer recorded values.

Predictive power

The ability to show customers what they will lose if they don’t act in addition to what they might save is a truly powerful tool. Hundreds of repeat cleans on the same vessels over years have provided the biofouling growth rates associated with antifouling coating products, technology types, and ages. Additionally, recent data exploration isolated seasonally specific growth rate pressure in Norwegian waters to allow for highly accurate modelling over a vessel’s dry dock cycle. 

With fouling density tied to a fuel impact value, we can model cost saving scenarios for various cleaning regimes which are vessel specific, factoring in antifouling coating type, degrading antifouling performance and seasonal growth fluctuations. A previously silent drain of resources is getting louder thanks to this new data insight that we can model.

Making time to answer questions 

Now, this data has always existed, but it took the right people to have the time, energy, and will to explore the potential. In a world of busy plate spinning, it took finding that moment – combined with the buy-in of the customer who we share vessel data with – to allow me to answer some of the questions that I’ve had for months (if not years). It also gave me the confidence to know that there was a lot more value in the existing data that maritime has been ignoring due to a lack of time. 

An unexpected benefit was not just the ability to answer new questions in real time, but the chance to backward engineer questions that I didn’t realise I could ask of the data. The hours I spent looking through the AI results may have been a smidge longer than what was strictly needed, but detailed human examination to confirm validity of data trends is important since many of the data deep dives were prompted by having sufficient experience to know what would be a useful question to ask.

I hope that more people use AI to cut the legwork and free up head space to think about the bigger picture of what the data may reveal.

———————————

Click here to sign up for our free monthly ‘De-Mystifying Digitalisation and AI in Maritime’ newsletter and get the latest cross-sector expert insights, news, interviews, videos and case studies straight to your inbox.