ABS Pioneers Condition-Based Model for Class

ABS is conducting a groundbreaking pilot project with the U.S. Navy's Military Sealift Command and working with others in the marine and offshore sectors to use condition monitoring techniques and "big data" for purposes of classification. Those partnerships are now bearing fruit in the form of new guidance for classing vessels based on their current and predicted condition, as determined by sensor data, inspection results and sophisticated analytics. If successful, this new class paradigm could reduce costs for shipowners while giving them more visibility into the performance of their assets.

Derek Novak, ABS’ senior vice president for engineering and technology, recently spoke with MarEx about the project’s progress and its potential benefits for shipowners.

MarEx: What is condition based class, and what does it look like today?

Novak: Condition-based class involves a classification model based on the actual condition of the asset, not a prescriptive, calendar-driven requirement. It relies in part on condition monitoring technology as well as newer digital and analytics based approaches to analyzing operational data to assess the asset’s health and inform maintenance decisions.

Our ongoing partnership with Military Sealift Command is the first condition-based class trial in the industry. Their ships are government vessels, so they fall under a different regulatory regime, and they comply with class voluntarily. That makes it easier for us to explore condition-based class concepts with them and test some of the underlying technology. With MSC, we've learned a lot about how to use ship data and apply machine learning and AI to make predictions about how assets may behave in the future.

Based on that work, we've recently come out with a set of guidance notes for smart functionality. These are oriented at implementing this technology in a structured format for the commercial world, where there are strict regulatory requirements involved. We are partnering with companies like NOV, Wartsila and GE for this work - they've already invested in their equipment to be able to provide more information to their clients, and we're working with them to leverage that information to make better decisions about each machine's performance. We believe that this will help our vendor partners and shipowners achieve new operational efficiencies, reduce maintenance costs and experience a less intrusive classification process.

MarEx: How does condition-based class work when it comes to assessing the state of the hull?

Novak: Hull condition monitoring technology has been around for a while, but we can now take that data and apply it to a digital twin to predict how that structure may behave in the future. Traditional hull condition monitoring is about what is happening now, but we're taking that condition information and we're coupling it with AIS data and weather information to model the loads on the vessel. This will assist with better understanding exposure and damage accumulation and also making predictions about the health of the structure in the future. It also allows us to perform our structural surveys in a very targeted fashion: rather than spending a lot of time on hull structure that is in good condition and hasn't been loaded heavily throughout its life, we're able to direct our surveys where they are needed.

We're piloting this with MSC by fully instrumenting one of their vessels for hull condition monitoring.  

MarEx: How will this method of classing ships be introduced in the commercial fleet?

Novak: In the commercial world, this will happen through incremental steps. Large merchant ships are governed by SOLAS and by flag state regulations, and we'll need to work with those if we're going utilize such approaches and still meet the intent of any of the statutory survey requirements. We believe that you can use this technology in an incremental fashion in order to perform less intrusive surveys and make more informed decisions.

By sector, I think that these benefits will be coming first in offshore, where the vessels usually operate near one particular location, and for domestic vessels or vessels trading on an A-to-B route. Vessels on transoceanic services, with multiple port calls and changing itineraries, will probably have the largest hurdles when it comes to regulatory requirements.

MarEx: How have marine safety regulators responded to the concept of condition-based class?

Novak: The regulators I've spoken with all have a pretty open mind. From a safety standpoint, condition monitoring means that we're going to have more information about the assets and we'll know more about what's happening. At a high level, they've all been open to exploring how to demonstrate equivalencies through the use of technology.

MarEx: "Big Data" projects for ships often come with a startup cost. Do you expect that the underlying sensor and IT systems will become more accessible with time?

Novak: I think we're seeing more and more of these technologies in use in the shipping industry. I recently looked at a new container ship that had 2,800 sensors in its control systems alone, and they're pulling in about two gigabytes of data per day. I think that this technology is going to become more affordable, but as an initial step, we're trying to help owners leverage what they already have on board.

The first question we get from shipowners about smart functionality is that they don't know where the first step should be. We tell them that the chances are that there's already a lot of technology on board their asset. Let's see how we can leverage that data to see if we can increase operational efficiencies, reduce maintenance, and reduce the intrusiveness of a class survey. Some owners may choose to add more instrumentation to their vessel, but that's going to be an owner-by-owner decision based on their strategy.

MarEx: How much customization is required to set up a new ship?

Novak: I usually break it down in thinking in terms of equipment (machinery and systems) and structures. Equipment manufacturers have a multitude of sensors already built into each of their systems. We can build a digital model for a given product, then use it as part of the model for each ship that has that product on board. Obviously, each ship will have unique aspects, but the equipment models are developed once and used multiple times. This is one of the benefits of our partnerships with some of the leading OEMs.

For structures, a jack-up rig is going to have a much different sensor layout from a container ship, but modeling differences between series of container ships should be relatively minor.

MarEx: How do condition monitoring and condition based class figure into the development of autonomous vessels?

Novak: To go fully autonomous, you really need to know your asset and understand what needs to be done to that vessel to maintain reliable operations. To accomplish that, you need to know what is happening on board from a vessel health and performance standpoint and take necessary corrective or preventative action before you send it back out to sea. Health and performance monitoring have a major role to play in giving the shipowner that knowledge.

There is a path to autonomy: you start where we are today, with no "smart functionality" and all decisions made by human crews. With smart ships, the decision-making is augmented using data analytics to assist or reduce crew. The next level after that is to let the decisions be made by the machine, but with human oversight, and remove the crew from the ship. The final level - full autonomy - means that the machine is making all decisions without human intervention. If we are going to operate fully autonomous ships, we will need to fully understand the performance, maintainability and health of each asset.

MarEx: Do you foresee a future when new ships are designed and built to enable condition-based class?

Many of the designers we're talking to today have already started this. Some of the major shipyards in Asia as well as some of the people here in the U.S. are looking at how to leverage the technology and how to offer shipowners and rig owners a way to start using their data. We hope to announce several projects with our partners in this area soon.   -- MarEx