FuelTrust.io is a startup with heritage in maritime and IBM’s Watson AI project, from the two founders Darren Shelton and Jonathan Arneault respectively.
The tech they’ve developed is an AI-powered ‘digital chemist’ that is able to crunch data on fuel profiles, proving that the fuel a vessel is about to burn is the same as what it thinks it bought. It provides rich data on the fuel, down to the batch level – which means that charterers, owners, and suppliers can reap the benefits and green premiums of supplying a slightly better blend of a given fuel – because not all batches are alike. This information is stored on a blockchain, which makes it more secure, and harder to introduce inconsistencies.
The ramifications of this are significant: it means that charterers and owners can work back through their historical consumption and set more accurate benchmarks for emissions, and there’s also a safety angle in terms of being able to tackle quality issues and fraud.
Shelton has 25 years in the maritime agency world, most recently at Moran, the largest private shipping agency in North America.
This is more like being able to get cleaner, cheaper gas by going to a different pump at the same station
A few years ago, he noticed that larger players were starting to lean towards blockchain for post trade finance settlements.As IMO 2020 came along, he started to see the opportunity – and need – to use this type of digital record to demonstrate compliance for shipowners in the fuels they were using, and that’s what eventually became FuelTrust.
Arneault, meanwhile, has 27 years’ experience in the software and high-tech spaces. For the last 15 of those, he was an engineer and then executive at IBM. FuelTrust is his fifth start-up.
“When Darren came to me, it wasn’t a case of ‘let’s start a company’ but ‘let’s solve a problem’. When we looked at the fuel supply chain, we saw that there was a range of bigger issues, quality, traceability, a significant amount of fraud in the market and the finance issues that underpin that,” Arneault recounts.
The upshot of this was FuelTrust, a risk provenance platform, tracing and validating the provenance of fuels and lubricants from earth to atmosphere, tracing the lifecycle of the constituent parts of a fuel across the entire chain.
“We are not simply documenting that a fuel moved from step to step – we are looking at molecules at each step and running the observations through our AI-driven digital chemist to work out what has changed,” Arneault says.
For example, has the sulphur changed because of evaporation, was the blend not correct, was it put into a barge that wasn’t cleaned?
FuelTrust collects transactional data from all the documents, and then combines them in an AI model. This AI does chemical analysis in a digital twin – essentially simulating reactions on a molecular level to reverse engineer the data and see what’s happened.
“One thing that’s surprising is just how much variation there is between individual batches of fuels,” Arneault says.
Between batches from the same supplier, FuelTrust has come across differences in energy intensity of up to 3%.
“This has huge ramifications,” Arneault insists. “In the world of zero-carbon fuels, we know that energy density is going to be a much bigger issue, with methanol, for example having less than half the density of HFO. So, people are going to be looking a lot more closely at the energy they are buying for their money.”
This is also likely to lead to big differences in emissions profiles and carbon intensity.
“In a world of emissions trading schemes and carbon taxes, you’re going to want to have better data on the fuel you’re burning – if the emissions turn out to be lower than the estimate for that grade, you’ll want to capitalise on that green premium,” Shelton says.
“To use a rough analogy,” Arneault explains, “it’s like we already know that you can get cheaper gas at a different gas station. This is more like being able to get cleaner, cheaper gas by going to a different pump at the same station.”
In addition to the discrepancies between batches, there are a lot of unknowns about current fuels hidden in the margins. For example, a recent study by FuelTrust identified an up to 0.3% difference between the stated density of fuel on a bunker delivery note (BDN) and the density of samples of the delivered fuel subsequently tested in the laboratory under the same conditions. Under FuelTrust’s study, 66% of VLSFO samples tested in a laboratory had lower densities than that declared on the BDN.
“The key causes of this bunker discrepancy are fuel temperature and volume at time of delivery, fuel moisture content, and more disturbingly, contaminants, which can be added at any point in the supply chain,” Shelton says.
