Trust Data, not dime-a-dozen opinions

Innovative-shipping-article-header-banner.jpg
If data truly is the new oil, as mathematician Clive Humby observed, the companies that are able to use it in the best possible way have a massive advantage.

Marine fuel price volatility and advancing environmental regulation have been major factors in many operational and investment decisions across global shipping in the first part of this century. However, recent rocketing bunker prices, ship efficiency and carbon intensity measures due in force from 2023, and an IMO target to reduce CO2 emissions by 40% by 2030 put the twin issues head and shoulders above any other cost considerations.

Assuming that the case for sustainability is compelling and accepting that soaring fuel prices create irresistible efficiency imperatives, ship owners and managers can start to feel coerced - rather than persuaded – to respond.

Given that the challenge in question exists now - in ‘real-time’ - responses that rely on immature alternative fuels or commercially unsustainable energy storage technologies are broadly beside the point, no matter how promising. Instead, shipping is better served by working on with solutions which optimize a global industry whose existing assets and technologies simply cannot be supplanted within the current GHG regulatory framework for change.

Answers lie in the myriad of real-time indicators already at work on board ships worldwide, which measure the performance of ship structures, ship machinery, engine torque, fluid flow and navigational changes. The challenge is to recover actionable intelligence from the gains hidden inside these complex and sometimes seemingly indecipherable data.

The volume of data concerned is so large that specialized techniques and methodologies are needed to manage, process and utilize them. However, while new to shipping, these techniques are also proven.

For the uninitiated, Artificial Intelligence and its component part Machine Learning are explained here as a huge set of algorithms, tools and techniques which are optimized to handle Big Data in ways that mimic human learning at multiple levels – and at scale. The process uses existing datasets to develop models which learn and easily recognize trends and patterns which would not be discernable to the human mind - or other mathematical methods - to predict future behaviors. Repetition allows models to train themselves and self-improve their accuracy and intelligence.

But AI is also ‘only’ a technology, which in a few years will be embedded on all household appliances. Therefore, before another new “groundbreaking” technology appears on the horizon to shake up the status quo, perhaps its emergence can provide the basis for a more lasting observation: the value of whatever that next technology is will surely be determined by its ability to make best use of and decisions about data.

After all, we already know that 24/7 data acquisition from all available sources provides a rich resource to develop a high-resolution and realistic picture of ship performance that is far superior to the patchwork of noon reports, manual datasets and standalone systems used today. The benefits of real-time data gathering and processing adds up to more than the sum of its individual technical, operational, and cost gains, and will depend on the user’s aims.

It takes little imagination to understand the significance of being able to use data on required speed, anticipated weather conditions, actual engine efficiency, hull and propeller cleanliness, etc., to predict ‘future behaviors’ such as a specific ship’s fuel consumption during its next voyage. Machine Learning transforms these rich datasets to competitive advantage – for example establishing a direct correlation between shaft power and fuel consumption for an individual ship which is not described by any mathematical engineering formula.

Monitoring for conformity to a regulatory framework also becomes straightforward. For example, METIS reports recommending operational adjustments in 2021 caused a tanker’s CII rating to fall from 3.8 points to 2.6 points without drydocking or hull cleaning. The ship is shaping up to save almost $1 million in fuel costs for the whole of 2022.

Without relevant data, all we’re left with are opinions and gut feelings.

METIS Cyberspace Technology is an innovative ambassador for implementing Machine Learning in the shipping industry. Its objective is to ensure an integrated and reliable process of data collection, real-time performance monitoring and intelligent analysis, providing actionable information for shipping companies. Using Machine Learning, the METIS platform evaluates itself every few days and retrains monthly to fine-tune the correlation between weather, hull fouling, shaft power use, fuel consumption and efficiency, and many other significant parameters.