Machine intelligence is a game changer in many businesses; can the shipping industry harness its potential?
Artificial Intelligence (AI) and Machine Learning (ML) are the two terms dominating the narrative in today’s business world. CEOs of the world’s top 10 companies swear by it as the most defining technology of our era.
The titans of FAANGS (Facebook, Amazon, Apple, Netflix, and Google) are personally visiting the big names in Machine Learning and Deep Learning and hiring them.
Microsoft CEO, Satya Nadela calls AI the new runtime engine of the company. Elon Musk recently described in one of his interviews that if there is one new technology he would like to work on, it would be AI. Sundar Pichai of Google claims that AI is at the core of their operations, and they are primarily going to be an AI-driven company.
In an AI driven company, there are no workers in the “critical path” of operating activities. AI runs the show. For example, in an AI driven financial operation, there is no employee providing financial services, there is no manager approving the loans, no employees deciding your monthly premium. Similarly, in AI driven advertising business such as Facebook, when you want to run an advertisement, there is no employee deciding the rate. With Alibaba’s Alipay when you want to send money, the entire authorization and transaction is driven by AI. Without the operating constraints of a traditional firm these companies can experience unbridled growth and scale their operations without increasing the cost proportionally. The age of AI is being ushered in by the emergence of new kinds of firms.
Google, Facebook, Netflix, Alibaba, Tencents and even many smaller companies are placing AI at the core of many of its businesses. Every time you use a service from one of those companies a completely algorithm driven decision making process takes place. Rather than relying on traditional business processes operated by employees, supervisors, managers, process engineers or a customer representative, the value and experience received is served by algorithms.
Of course, the algorithms are written by data scientists and engineers, the process is set by process managers, the front end UIs are designed and coded by engineers but after that, the system takes over and delivers the value on its own. AI sets the prices on Amazon, recognizes your speech by adapting to your accent on Google, rearranges photos in separate albums in Google photos, recommends songs in Spotify, and qualifies borrowers for an Ant Financial loan.
At the core of such companies are decision systems. It treats decision making as a science. It analyses internal and external data and converts them into predictions, insights and choices which in turn drives the automated business workflows. Its algorithms decide what price you pay to Facebook for placing an advertisement, run ad-auctions on Google, decide which car you get on Uber, Lyft and Grab. It sets prices of goods on Amazon and even runs the robots that clean airports at Airports and Walmart.
The traditional firms in the same business have no choice but to follow the trend if they want to stay competitive. Firms like Walmart, Honeywell, Comcast, Fidelity, Charles Schwab are now turning extensively to data, algorithms and digital transformation to effectively compete and remain relevant in their businesses.
When we all hear the term AI, thanks to Hollywood, it brings an image of highly sophisticated machines or systems, indistinguishable from human traits and simulating human reasoning. Some academics including Dr Lakhani and his cohorts call it “Strong AI”, probably to distinguish it from what’s possible today. What we are talking about here, along with the various examples of AI usage above is simply described as “Weak AI”.
Weak artificial intelligence (AI)—also called narrow AI—is a type of artificial intelligence that is limited to a specific or narrow area.
Weak AI simulates human cognition. It has the potential to benefit society by automating time-consuming tasks and by analyzing data in ways that humans sometimes can’t. Weak AI can be contrasted to strong AI, a theoretical form of machine intelligence that is equal to human intelligence.
Weak AI helps manage the information business such as Google, Facebook, and Twitter. It also delivers services in the entertainment business such as Netflix and Spotify.
There are certain use cases in the early adoption phase in which AI will have a really large societal impact. It will guide how the company builds, delivers, and operates the actual physical products. Case in point, Amazon’s warehouse robots, Brain Corp’s floor cleaning robots operating at Walmart’s, or self-driving cars proposed by various firms such as Intel’s MobileEye, Google’s Waymo or Uber’s Rideshare.
A semi-automated process that gathers, cleans, structures, and integrates the data in a systematic, sustainable, secure, and scalable way.
Using both internal and external data, algorithms generate predictions about future states or actions of the business.
The systems that embed this entire process in software and connect it to internal and external users through a user-friendly interface.
Given the nature of the shipping business, we see humans in critical positions for the foreseeable future; however, there are various functions, tasks, and subtasks that can be made much more efficient by employing AI at the heart of the operation. This can have significant ramifications in productivity gain, operational efficiency, and reduction in cost.
This has been a domain of expert navigators who use their knowledge to make a call on route and speed. However, humans can’t match the AI-enabled optimizations as the machines can run millions of permutations in minutes to calculate the most economical speed, given various constraints such as weather, waves, current, laycan etc.
For example, Alpha Ori Technologies’ AI platform can run 80 million permutations and combinations in a matter of minutes to identify the most economical speed considering all the relevant variables – hull design and condition, load, weather, waves, current, laycan etc and provide recommendations for best fuel savings, best TCE (Total Charter Equivalent) or a hybrid result without any human intervention. Since 2018, the company has been experimenting with AI tech on various types of cargo vessels (Bulker, LNG, LPG, Tankers) where it has proven that AOT’s AI software can provide fuel savings of up to 10%. Customers have taken notice and AOT has been winning fleet wide expansion contracts lately. The below chart provides a real-life example of potential savings as well as actual savings, realized in a large-scale study conducted on over 60+ vessels in mid 2021.
AOT, in partnership with a ship owner, did an analysis to identify the impact of the Total Fuel Oil Consumption (TFOC) module on the fuel savings potential for a specific vessel. This vessel was a bulk carrier traveling from Yeosu to Port Hedland in the month of November-2021. Below are the details of the voyage.
Vessel Name | Iron Miracle |
Vessel Type | Bulk Carrier |
Gross Tonnage | 92379 |
Loading Condition | Ballast |
Selected Benchmark | Charter Party |
Start Port – End Port | Yeosu to Port Hedland |
Departure Time | 2021-11-02 9:42 |
Arrival Time | 2021-11-13 23:00 |
Voyage Duration (days) as per ETA/ETD | 11.6 |
Vessel status | Ballast |
Displacement | 92000 |
Average Speed Required | 12.8 |
Voyage Metrics | SMARTShip™ |
Total Ground distance (NM) | 3523 |
Total ME Fuel Consumption (MT) | 392.9 |
Average SOG (Knots) | 12.7 |
No of Good Weather days | 10 |
TFOC Metrics | SOG (Knots) | M/E Consumption per day (t) | M/E Consumption (t) |
CP Benchmark (11.5 days) | 12.8 | 37.8 | 433.5 |
SMARTShip™ Predicted | 12.8 | 377.1 | |
Ship’s Reported | 12.7 | 34.0 | 392.9 |
The above case study highlights the direct impact and benefits of implementing the SMARTShip™ TFOC (Total Fuel Oil Consumption) solution on cargo vessels.
The average total fuel consumption for the charter party benchmark for an 11.5 day voyage was 433.5 (t) versus the reported consumption for one of the vessels which used SMARTShip™ recommendations was 392.9 (t). This was a straight-up 9.4 % fuel saving against the benchmark, with additional CO₂ savings contributing towards decarbonization goals.
AI can be used to analyze multiple navigation scenarios. Orca AI, an AI navigation platform, is planning to provide such a solution by combining sensors and cameras with deep learning algorithms. Alpha Ori is in discussion with a US based company, Everguard, to jointly work on a product module focusing on both safe navigation and safety onboard during the voyage as well as cargo loading/ unloading in the port.
The goal of the proposed AR navigation system, an advanced augmented reality navigation system is to provide substantial support to navigation, using the power of augmented reality.
With a video camera pointed in front of the vessel, the front view image is projected on a display and all the necessary navigation information is superimposed over this live video imagery by the AR technology. One can clearly view other vessels’ routes and critical information, as well as own ship data even in adverse weather or visibility conditions, allowing safe maneuvering and navigation. The augmented display and sharing of information provide enhanced situational awareness, crew confidence, watchman support, and allow for better coordination of crew members.
The project aims to contribute to the safety and security of the voyage by offering visual support for maneuvering and navigation during any operation.
Predictive maintenance using AI / ML models has been successfully deployed in other industries such as manufacturing plants, aircraft maintenance and high-end military equipment maintenance. It’s the right time to bring the technology to the Maritime industry for identifying anomalies and predicting failures. AI models can help maintenance managers predict the likelihood of future failures and determine asset failure factors that could impact ship operations. If done correctly it has the potential to
The below case study is another great example of how SMARTShip™’s Asset AI solution uses features like dynamic alert mechanism, advanced anomaly detection, calculation of asset health and provides remaining useful life (RUL) estimations which result in direct cost avoidance or savings for the maritime industry customers.
In November of 2021, a vessel’s auxiliary engine was exhibiting low turbocharger speed resulting in low scavenge air pressure and increased exhaust gas temperatures across the engine.
Figure 1 displays the Asset AI chart of turbocharger speed vs power.
The chart displays a baseline curve and warning/alarm offset curves of turbocharger speed relative to generator power and, overlays the most recent data for comparison. The baseline curve and alert limits are generated automatically using machine learning models based on historical operating data (first 1,000 equipment operating hours). From a quick glance it is clear that there has been a significant reduction of approximately 2,000 rpm.
In combination with this parameter, Figure 2 displays the Asset AI chart of turbocharger outlet temperature vs power.
As can be seen in figure 2, it is clear that in parallel to the turbocharger speed reduction, there has been a significant increase of approximately 65°C in the turbocharger outlet temperature.
Asset AI is able to identify all the different anomalous parameters occurring together to provide users a ranked list of probable causes. Further, the application stores maintenance history and collects engine run hour data, allowing for the estimation of remaining useful life of components by assessing their age and current performance.
In this case, Asset AI identified the poor engine performance early and was able to properly identify the root cause which was later corroborated by the ship’s crew after inspection. The timely alert was received by the ship’s crew who inspected the turbocharger and found significant deposits. The turbocharger was dismantled and cleaned of deposits on the nozzle rings and rotor and put back into operation. Following the maintenance, all parameters were observed to have shifted back into the normal operation range. If maintenance had not been performed, it is probable that the turbocharger would have been irreparably damaged.
This early detection resulted in an estimated savings of approximately $40,000 (cost of T/C cartridge renewal), as well as additional savings due to improved fuel consumption and extended life of engine components that would have been adversely affected by the poor turbocharger efficiency.
Asset AI provides a unique approach to AI; for more information click here
Despite the considerable efforts of the maritime authorities, safety is still a concern especially in areas of heavy traffic. The European Maritime Safety Agency (EMSA) has reported 20,616 marine casualties and incidents worldwide from 2011 to 2017. Globally, the combination of collision (23.2%), contact (16.3%), and grounding/stranding (16.6%) shows that navigational casualties represent 53.1% of all casualties with ships. Furthermore, human error is found to be behind 75% of 15,000 marine liability insurance industry claims analyzed by Allianz Global Corporate & Specialty (AGCS).
Collision avoidance can make a significant impact in reducing collision and contact at sea. It is a safety system designed to warn, alert, or assist bridge officers to avoid imminent collisions and reduce the risk of incidents. Collision avoidance systems use a variety of technologies and sensors, such as radar, lasers, cameras, GPS, AIS and artificial intelligence.
AI and deep-learning technology can be used for ship image recognition systems very effectively. We are all aware of the exploits of Facebook and Google in using AI in face recognition and image recognition technologies with enhanced accuracy. Similar image recognition technology can help the maritime industry in avoiding collisions that end up costing millions of dollars for ship owners and P&I clubs.
Such a system can use ultra-high-resolution cameras and a graphic processing unit (GPU) to automatically identify vessels in the surrounding area. An AI enabled app trained on a large set of floating object databases, can provide a visual reference depicting where a target or multiple targets are on a radar-range-like graphical screen, including the target’s speed, bearing and closest-point-of-approach data. Additionally, the app can trigger onboard alarms and warn of detected targets. It can help improve safety and help stop large vessels colliding with smaller ones, especially in and around ports. It can also provide alerts to other hazards, particularly when visibility is poor. The image recognition technology could be used to monitor shipping lanes, as well as for security and coastguard operations.
Theoretically, it increases or decreases based on the risk estimates made by the insurance provider based on the following.
However, many will tell you that the final premium is determined by the back room deals, relationships and number of vessels offered. In this day and age, it can be considered archaic. Nowadays, the ship’s operational data is available thanks to digitalization. An AI algorithm can assign a risk score monitoring ship’s live data from sensors and ERP systems, various inspection reports, maintenance records, crew nationality and many other factors mentioned above. The risk score is dynamic and transparent which incentivizes ship owners to maintain their vessels for lower premiums.
Obviously, these are just a few examples of various operations where artificial intelligence can play a game changing role in making shipping cost effective, safe and sustainable.
Given the sheer size, scale and importance of shipping as a driver of global commerce across all industries, the visionary leaders of the industry are ready to bring in the digital technologies to disrupt the status quo. At the same time, the new technologies (AI, ML, Cloud Computing) have reached a level of maturity that they can be implemented at reasonable cost without much risk.
Lastly, we all know the headwind this industry is facing – Climate change, rising manpower cost, marine pollution and demand for safe navigation etc. It’s a matter of time before shipping companies are driven to accept newer technologies such as artificial intelligence that are already proven to benefit other industries, and hold the promise to transform maritime transportation.
Sam Jha is the Chief Product Officer at Alpha Ori Technologies (AOT). Alpha Ori is a B2B Technology company operating in IoT (Internet of Things), Ship ERP (Enterprise Resource Planning) and Big Data Science. For any comments reach out to sam@alphaori.sg