The five hidden inefficiencies machine learning can help shipping companies find

The shipping sector has a history of conservatism, but the need to embrace change and make use of modern technology is ensuring a widespread change in mentality. The combination of big data and machine learning is helping the industry work out where it has been going wrong, and discover how it can become more sustainable.

Chris Pontet

30 Apr 2019

Shipping is the backbone of global trade, with some 90 percent of goods transported by ship at some point. This degree of importance may go a long way to explaining why shipping inefficiencies are ingrained throughout the sector: out-of-date technology, high numbers of manual interactions, and traditions and culture outweigh a drive for embracing the new.

Whether ro-ro, gas carrier or cruise ship, gains small and large are available to any vessel actively seeking ways to operate smarter. By saving on fuel, optimizing capacity usage, planning more efficient routes, reducing waiting times, planning maintenance more effectively, and operating more efficiently—all of which ultimately reduces costs—, companies can thrive in a period of major transition.

Of course, changing times means that evolution is key. Failure to adapt is liable to be costly, and, as with every business decision made regardless of industry or sector, it always comes back to money. Inefficient spending and poor investment choices can cripple a business, especially at a time when tight profit margins are set to be squeezed further.

Quite simply, the time for complacency in the shipping sector is over.

And, with that in mind, here are five lesser-known ways that big data and machine learning can help you and your vessels save money.Wasted-fuel-and-harmful-emissions-[Insert]

1. Wasted fuel and harmful emissions

Emissions from the shipping sector account for more than 10 percent of the total carbon dioxide emissions from global transportation. In a world where fuel costs are rising, and global emissions standards are increasing in stringency—the 2020 IMO fuel sulphur regulations, for example—, reducing this is vital if operators are to survive.

One huge step the sector could take is to phase out the use of noon reports for information on fuel efficiency. Data on vessel speed, weather, temperatures, pressures, heading and draft is available every few minutes on the vast majority of vessels, so why not make better use of it?

Instead of relying on a data-check once a day, all vessels should be moving towards an automated system that cuts out manual data entry, and takes account of the constant shifts in a vessel’s real-time operating performance and status. Not only would such an approach incorporate minute shifts that could have a huge impact over time on fuel consumption, but it could also be used to predict future events.

Being able to acknowledge vessel performance minute by minute is, quite clearly, a more efficient means of operating. By affording operators a level of oversight not previously possible, better decisions can be made exactly when they will be most beneficial.Unplanned-maintenance-[Insert]

2. Unplanned maintenance

Mechanical and technological failures are the root causes of a high proportion of delayed deliveries, and also one of the highest costs to shipping. The automatic recording of the wealth of performance data from all parts of the ship and its analysis by artificial intelligence enables the use of predictive maintenance. This means operators can assess individual pieces of equipment and their behavior in relation to others, and can then schedule maintenance prior to a problem arising.

Machine learning below the waterline matters, too. Hull fouling increases resistance, slows vessels, significantly reduces fuel efficiency, and raises fuel costs. Machine learning can help predict the degree of fouling over time on all hull types and according to underwater conditions, and can flag up when it’s time to either clean or entirely recoat the hull.

Maintenance planning with the aid of machine learning means shipping companies can move away from approaches that are either overly reactive or overly timetabled, allowing them to embrace a predictive and proactive approach based on better a understanding of why equipment failure occurs.

Advanced algorithms inform the vessel’s managers that an element of equipment requires repair long before humans or a logbook would be able to ascertain the same thing. Not only can this save shipping companies time and money, it reduces the likelihood of encountering unexpected breakdowns, and enables the provision of a more reliable service to customers.Anchorage-delays-and-unplanned-idling-[Insert]

3. Anchorage delays and unplanned idling

How much time do shipping operators spent receiving, and replying to, emails, telephone calls, and even faxes as part of the process of preparing a ship to enter port? Vessels, port authorities, terminal operators, agents and port services providers all scurry around, attempting to make the necessary arrangements, and still it’s far from uncommon for a vessel to steam ahead, eating up fuel, only to approach the port and find that another vessel is occupying its berth.

As well as the fuel inefficiencies associated with having to remain idle while awaiting a berth, this also results in delays, which will have knock-on effects at other ports which will affect customers and the wider supply chain. Linked up communications and machine learning could be used to ‘read’ the berthing conditions at the destination port, calculate likely delays, and enable adequate plans to be made. Operators would then be able to inform customers and other parties in the chain much more proactively and reliably, and reschedule in good time if required. Half-empty-vessels-[Insert]

4. Half-empty vessels and half-wasted voyages

A half-laden vessel is like half-loading a washing machine – you only achieve half of the potential gains, but with the use same amount of energy. It is even conceivable that, in future bids to curb harmful emissions, vessels will only be allowed to enter a port when it is at full capacity. That might sound far-fetched, but with shipping accounting for such a significant portion of the world’s global carbon dioxide emissions, it makes sense to seek efficiencies wherever possible.

By utilizing data and machine learning to better understand trends associated with transactions, and then predict demand and capacity usage, shipping companies can produce a range of cargo configuration options that will enable operators to more easily match smaller cargos from different customers, thereby enabling them to plan efficient capacities.

Shipping could learn a lot about the optimization of cargo planning from the road haulage sector. The trucking industry is making extensive use of algorithms to calculate pricing, plan routes, and match loads to vehicle/driver. Taking into account real-time events, trucking firms can better match cargo-loads to vehicle space, maximizing cargo density for each vehicle, and minimizing cost per cargo unit.

It’s easy to see why such strategies will allow companies to be more competitive and sustainable over an extended period of time.Communication-and-a-lack-of-understanding-[Insert]

5. Communication and a lack of understanding

Booked vessels that don’t show are a major inefficiency. Often, such events are the result of customers becoming anxious about tight availability, subsequently overbooking, and ultimately going with another supplier.

Because the shipping industry—particularly the container sector—generally does not charge customers for no-shows, costs can be significant, which will inevitably hit voyage profitability. However, at the same time, careful planning for efficient loading and operation—both for vessels that show and those that don’t—eats up time.

Sophisticated data analysis can help operators to detect bookings made by ‘unreliable’ customers, and uncover patterns that allow them to gauge the likelihood of no-shows. This enables pre-emptive action to maximize the chance of having a ship that is operational at all possible times, therefore making full profit.

Data and machine learning can also be employed to assist in the consideration and settlement of vessel performance claims from the charterer, which is a common issue in the case of time charter contracts. Accurate data analysis can help with the determination of speed, fuel consumption, weather, and numerous other factors that affect a ship’s efficiency, which can in turn help resolve discrepancies and disagreements that, traditionally, could have ended up costing the owner.New call-to-action


Book a demo

Sign up now