20 minutes to read Today’s podcast guest is Kevin Daffey, Vice President Governmental Engineering and Marine Automation at Rolls-Royce Business Unit Power Systems. In this episode of the Data Today podcast, we discuss Kevin’s impressive career, the vast flows of information that are required to keep the marine industry moving, and how shipping can reach net zero. Podcast transcript Dan Klein: Hello, and welcome to Data Today, brought to you by Zühlke. I'm your host Dan Klein, and I look after everything data and AI at Zühlke. We're living in a world of opportunities, but to fully realize them, we have to reshape the way we innovate. We need to stop siloing data, ringfencing knowledge, and looking at traditional value chains. And that's what this podcast is about. We're taking a look at data outside the box to see how amazing individuals from disparate fields and industries are transforming the way they work with data, the challenges they're overcoming, and what we can all learn from them. Everything we do in the modern world creates data and nothing creates data more so than international trade. At the center of that lies shipping. Our guest today is at the heart of this. Kevin Daffey is Chief Engineer at Rolls-Royce Marine Division. There are 60,000 commercial ships in the world, forming a complex network of trade routes, but shipping presents whole load of challenges. It's polluting and relies on extremes of human skill and engineering in all sorts of weather conditions. It's also very, very complicated, crossing over international borders and traversing great swathes of sometimes very conflicting legislation. And that's before we talk about the vast flows of information that keep these ships moving and the shipping lanes open. Where did you start? What got you into marine engineering? Kevin Daffey: Yeah, good question. Do I remember? Let's go back in time. So the first 10 years of my life, I spent my time in the steel rolling mill industry doing automation and electrical and drive systems. Really, really, really enjoyed that business. Watching hot steel rolling, and controlling it, it was a sort of a multi-variable problem, controlling temperature, width, thickness, shape, profile, with actuators and things like that. And you're trying to roll lots of steel, at various different grades as quickly as you can, consuming huge amounts of power. I did that predominantly as an engineer and a team leader of a small team, and then I got into engineering management, but mainly with variable speed drives, drive systems. So my background is electrical drive systems, software, controls, and then the Type 45 destroyer program in the UK kicked off around about 2000. And the company I worked for did the complete integrated electric propulsion system and I was plonked in as the chief engineer. And that's where I started in marine. So around about 2001, I started in marine industry as the chief engineer for that particular program. Really, really enjoyed that program, building 20 megawatt ocean units and power electronics and drives. And then in 2003, Rolls-Royce made me an offer I couldn't understand, but I decided to join them and help them on their journey to, in their marine business, build more electrical capability. And I did various roles. We bought a company called VT Controls down in Portsmouth, I worked down there for a couple of years. I had a really interesting three years on secondment into the UK Ministry of Defense, into defense equipment and support, between 2008-2011, I was Chief Marine Engineer, Electrical, I was also a naval authority, so I was there to sign off the power and propulsion and steering, or maneuvering systems, as they were called, for various ships. And I was involved with Type 45 again. The first three ships, I was involved with sea trials and permitting them to do their sea trials and then signing them off for the Royal Navy to say that they were safe to go to sea from the perspective of their portion of maneuvering systems. And played around with some marines and provide technical advice. And then for four years, I wasn't involved exclusively in marine, I was helping Rolls-Royce develop their electrical capability and I was working with university technology centers. I was ending up as director of five of them. And we were developing electrical technology mainly for aerospace applications. Did some marine stuff, working on power electronics and battery systems for marine and supporting our marine division in some of their product developments. 2016, I went across to Rolls-Royce Marine Division and was Director of Engineering and Technology. And in the middle of that I took on the role of running the, what was called ship intelligence, which was looking at autonomy and data. And that's actually where we met, Dan, was... Dan Klein: Indeed, in Norway. Kevin Daffey: In Alesund, I think it was. Dan Klein: Yeah, yeah. So I'm interested, if you scroll forward to 2022, in your mind, what's the emergent challenges, if you like, for marine now? I mean, where are the challenges for you now? Kevin Daffey: Coming out with COVID, there was, all of a sudden, a massive change that happened in marine in terms of a focus on sustainability and how does the marine industry meet the challenges around getting towards a net-zero industry by 2050. The IMO really pushed forward with some hard-hitting requirements. It started looking at the regulatory regime. So people are looking at what is it that we can do to get there, but it's not going to be feasible to bring every ship alongside and change its power plant and put in some super-duper power plant that will be net zero. That's going to take a long time. And we don't even know what the technology and the fuel is that's going to do that. There's a whole range of different solutions there. So one of the things that can be done in the interim is actually how do we optimize the way that we utilize ships? There's a lot of data onboard ships. And we know that the way ships are sailed, especially cargo vessels and dry cargo vessels and oil-carrying vessels, they charge off at 19 knots and they arrive at ports, and they stick around outside of ports and then they dock and unload. There's nothing there that takes into account that if you slow down a ship, you use a vast amount less power. So if you drop by a couple of knots, you could drop 30% of the power requirement, and therefore you drop emissions, you save fuel. Then there's no optimization at the moment of the way that we bring ships into ports, as it were. So if you think of the marine industry and shipping as a process, then there's a lot of process engineering work that could be done to optimize that. And I think we're a pretty chaotic, unoptimized industry that needs optimizing, and a data allows us to get there. You've got lots of external disturbances like the weather and tides and currents, and again, a lot of that is predictable. And forecasting is quite accurate for weather, winds are pretty predictable, tides and currents are very, very well known. So again, all of these provide you with marginal gains, your own power nature might be able to help you, for instance. So again, you've got this opportunity of taking this high-level view of optimizing everything. So any ship potentially could be optimized around its process effectiveness and how much fuel it consumes and how much CO2 it emits. Dan Klein: So now, in 2022, you're at MTU and you're based down in Friedrichshafen with Rolls-Royce. But you've described the power plants that MTU produce as the Ferrari of the marine world. All I can hear now is Jeremy Clarkson in my ear saying, "You've basically given Ferrari to all these boat captains and you're now going to say to them that they have to drive them like a G-Wiz." How well is that going to go down with these boat captains? Kevin Daffey: The high-speed engines that we design and develop are performance engines, but they're optimized for certain vessels. We use them in luxury yachts, for instance, production yachts and mega yachts, they're used in tugboats, they're used in work boats, crew boats, they're used in certain ferries, especially fast ferries, and they're used a hell of a lot in naval applications. Many of those applications, you need power density and you need performance. So for instance, a tug is a good example because tugs work in close proximity to ships which generally are larger than them. If you've got lines attached, it's very easy for the tug to end up being broached and turning over. And that happens. So you want really, really good maneuverability. When the captain puts his lever forward, he wants performance. So that's why you have the performance. But the engines are also, they can also be optimized to maximize efficiency. But yeah, you're right, there is a cultural aspect of how do you control the driver, the captain, the navigator, to then take a different approach. So fast ferries normally trying to meet a timetable. If you're trying to meet a timetable, you can't afford to have too many way of delays. So when we bring data and insights to the captain, we have to be conscious that we're providing advisory, "If you drive differently, you could save this." And it also requires the company that manage the vessels to maybe incentivize differently, to incentivize behaviors. Otherwise, the behavior will be, "I need to meet my timetable," because that might be the most critical KPI that that captain has to meet. The other thing you have to remember as well is that the waters in harbors and rivers is quite congested and there's a load of unpredictable things that can happen. Really, the information that we provide in terms of insights or where we're going with it is these are advisories, they're there to help you. Now, when you get to autonomous vessels, then those insights can become the commands, they can become the waypoints, the navigational points, go to that point at this speed, or remotely control vessels, where maybe there is a pilot on board, but they're hands-off and the system just drives them in that particular direction and does that command. Dan Klein: So it sounds like, in some senses, you've got a roadmap here from introducing advisory into boat captains, getting people comfortable with that technology, comfortable with how it works, stepping them into, say, remote controller vessels, and then the future is autonomous. How far away do you think we are from autonomous? Kevin Daffey: Yeah, good question. I think that we won't be too far away for certain applications. I think there'll be some ferry routes which will go autonomous. That's not to say there won't be anybody on board to take over, but it could be a little bit like Tesla cars where the driver's there, but everybody's hands-off, as it were. That technology, you can do that today. And that's been proven on various demonstrators. And I think that for autonomy, there are certain vessel types which will suit autonomous operation, and then there'll be other vessel types which really won't. So anything that goes on the big blue ocean and between countries, they tend to be bigger vessels. I don't think that we'll see too much in the way of autonomous control there because there really isn't a huge business case for it. And also I think there may be some political reasons why governments may not be happy with a ship that comes into their port being controlled by somebody who's in another country. So I've already had some of those conversations with one or two coastguards who say, "No, that's never going to happen." But you could see things like remotely controlled tugs being utilized, or pilot vessels, ferries where they're trying to optimize the route and make sure that the ferries dock well, especially if they're going to be docked precisely and then receive a recharge of electricity if they're battery-powered. So all of that might be done autonomously. And the other thing with battery power is that if you want to minimize the number of batteries on boards, you want to make sure that you can optimize the way that the vessel is driven, such that you can justify having the minimum number of batteries on board. And again, that's where this optimized navigation comes in because you can take into account tide and weather and currents, the passenger loading, the number of cars on board, you can take that all into account to really minimize the amount of discharge of the batteries. And then when you're loading back up again, with power to recharge it, you can, again, make sure that's as short as possible, but you get the right amount of charge in as well. So I think there, where you got a complexity of discharge and charge with batteries, I think that's a really good example of where some degree of autonomy would be really, really helpful. Dan Klein: What's taken the marine industry so long in terms of getting to this point? It seems to be a bit of a laggard. Is there reasons behind this? Kevin Daffey: There's a number of different reasons. First of all, every ship is a prototype, so every ship is different. The best way of describing it is a prototype. You might get a series of 2, 3, 5 or 10 of that boat type, ferry type, but that's it. Then the next one will be different. Also, the way that the whole industry is contracted, an owner will then commission a naval architect to design a ship, that will then be contracted to a shipyard to build it. The shipyard will say, "Right, I want this piece of equipment from there, this piece of equipment there." It's all optimized around the CapEx. Dan Klein: Doesn't that make retrofitting new tech and data quite difficult? And some of this data tech, I mean, you've talked about doing ship intelligence historically, you can't really retrofit the old vessels can you? Kevin Daffey: You can, but with difficulty. Also, if you look at data and controls, it's federated. You've got a federation of different control systems on vessels. None of them are necessarily well interconnected. So if you look at cars, there'll be a network that go around the car, a CAN bus link that connects everything together, and you can actually put a sniffer on your CAN bus and you can almost look at all the 3,000 parameters that are generated within the car. And there's even standards around the parameters. Same on an aircraft. Even though you may have some federated systems, there is some connectivity there that allows all of that data to brigade it together. There's nothing like that. There's no regulation in ships that say you have to connect all the systems together. And then the other thing is connectivity to ships is not particularly strong. So once you get ships going out into international waters, they then rely on satellite links. Satellite links are expensive, they're slow. Obviously that's low Earth orbit, satellites are beginning to change that paradigm. But in general, connectivity is weak. Dan Klein: I gather you have an experimental ferry at the moment in MTU. Are you able to describe some of the data sets you're looking at within that experimental ferry and what your thought processes are around how you envisage the future for that ferry? Kevin Daffey: So I wouldn't describe it as an experimental ferry, it's a ferry which transports passengers between April and October from Hamburg to Helgoland. And it's run with one of our customers FRS Ferries. So it's a high-speed ferry that's got four of our engines on, our Series 4000 engines on. Dan Klein: These are the Ferrari we've been talking about? Kevin Daffey: Yes. These chuck out a combined total of 8 megawatts out the back of the vessel through water jets and propel it at well over 30 knots, and moves very quickly through the rough the North Sea, as it were. So our customer there is basically allowing us to help them optimize how they operate and maintain the vessel by us instrumenting up the engines and the propulsion system. And we've been using it as a way of developing additional functionality that can be used for both advisories on board the vessel as well as ultimately we want to bring it out to fleet level so they can look at this ferry alongside other ferries in their fleet, and they can get some quantitative view and insights into their operation. For them, they're looking to how do they reduce the cost of the fuel that they use? Because obviously, passengers pay money, that's their main revenue source, a bit of money is made on board from selling them cups of coffee and food. And then obviously on the negative side of the P&L is the cost of the crew and the cost of the fuel and the cost of the maintenance. And they don't have a huge maintenance window. The vessel comes alongside at eight o'clock in the evening, it has to be readied then for seven, eight o'clock in the morning before it goes out at nine o'clock to take passengers out again. So you've only got a short window to do any maintenance. And another factor is that they can operate on three engines. If they lose another engine, then obviously the ship can't sail. So that's a loss of revenue for that particular day. So again, availability's key, they've got to keep the revenue ticking over. So we're looking at how we can use data to both prognose issues with the power plant, not only with the engines, but also with the propulsion plant by looking at bearings and the water jets and gearboxes, as well as looking at what fuel do they burn, what CO2 do they emit, looking at the journey that they take, looking at the speeds they take and really beginning to grow some insights into it. Dan Klein: When Kevin began his career at Rolls-Royce, the use of data was limited in marine. Kevin has been key in making its applications to ships up to date. Achieving net zero by 2050 is going to be a huge challenge for this industry, but it seems possible because we simply have so much data to work with. It's true that the marine industry has made great strides in their mission to modernize, developing autonomous technologies like they have with self-driving cars, and implementing satellite connections like those used in the aerospace industry. Data collected by ships gives the industry an idea of how to continue formulating and optimizing these methods. However, additional factors such as boat captains, passengers, the weather, tides, sailing times, and fuel costs, they've all got to be taken into account when formulating new methods of effectiveness, making it an extremely complex process. So how does the industry use data sets to optimize ships' performance? Kevin Daffey: It's understanding your process. So if I go back to the steel mills, used to log a lot of data from the steel mills, you're looking at the quality of the yield, everything that comes out of it, and you are looking at, "Well, there's some sensitivity here, something's happening here that I don't understand, I need to understand it. And if I can overcome that, I'll increase my yield or reduce my energy." It's the same on a ship. It's really understanding as much of the context as possible. So you can get some insights by just looking at engine power, maybe exhaust temperature, and that will give you some idea about how well the engine's performing. But we want to look at, "Okay, you got four engines on board, is one engine taking more power than the other? Is that an engine problem or is that water jet issue? Is the water jet getting failed?" Over time, this ship is using more power to go at the same speed, but other things are not changing. So is the hull getting dirty? What's the cost of cleaning the hull? How long does that take? If I clean the hull and I take two days of revenue out to clean the hull, but I put it back into service, but I'm using less fuel, will I save money? And then if an engine breaks down, then obviously I'm going to lose that revenue for a day. So can I run this engine for another 100 hours before we need to make an interventional or just do I need to do something when I get back to port? So all of this is insights, which potentially the data could unlock. And ultimately, you're trying to take variability out of all of this and bring consistency to the operation. Dan Klein: Each of these small details in the data affect how decisions are made to maximize efficiency and performance and minimize emissions and cost. While ships are traveling, not only are they collecting highly valuable confidential data, but they can also hold protected information about their owners and operators. In the data world, legislation, national boundaries and regulation are big topics. Now that the marine industry is using the same satellite connections as aircraft, data is on the move across these legal boundaries. Questions begin to arise around data ownership and control. You can have a ship owned by one nation flagged to a different nation, and in international waters or territorial waters, depending on what it's doing. So who then owns and operates that data and insight and where's it stored? Are you able to give us a view as to some of the complexities that are involved in data being on a vessel and being moved off a vessel even when you've got a ship moving around the world? Kevin Daffey: Yeah, it's really hard to be general about it because you have to get into specifics. I mean, ultimately, the data's owned by the owner of the vessel who's operating it. It's their ship, they pay for it, all the equipment on board and any data it generates is theirs. If they pass it off the ship, then obviously they have to be cognizant of data protection laws, which can exist in different jurisdictions, certain technologies, and the data associated with them can also be subject to export control. OE equipment manufacturers might not like their data being shared with one of their competitors because there might be intellectual property that's embodied with it in it. So now you've got a really complex mixture of challenges around what you can do and who can see that data. So the techniques we can use is to use anonymity, just to say, "That's data from a ship. It might be a tugboat, but I don't know whose tug it is." So again, there are some strategies that you can employ. When it comes to naval vessels, you've now got national security that comes into play and who can see the data, especially if there's anything related to context like position, low profiles, there can be quite a lot of sensitivity around some of that data. So pragmatism can come into it, but that's not always the case. Hugely complex, you need a law degree to understand that, and it has to be an international law degree as well. It's something which the marine industry has to grapple with every day. But then you look at aerospace, the aerospace industry, that flies from country to country, OE manufacturers, aerospace IP is well protected. That seems to manage it, so what can we learn from aerospace? Dan Klein: Just like with the automotive industry, we're now starting to reckon with the power of autonomous ships. Maybe we're looking at this issue in the wrong way, maybe it's more about responsibility than it is about opportunity. Politically, you might be in hot water if your captain is in France, but your ship is in the South China Seas doing laps around Palawan. Data is an extremely complex topic, whether it be analyzing it, storing it, protecting it, or moving it. If you've been listening to Kevin or any of our other guests, I'm sure you'll have worked out that we don't have all the answers yet, but experts like Kevin have the determination to find them. With data, he understands the importance of looking into every detail, recording every bit of information, and thinking of every possible scenario and outcome in order to use the data to its fullest extent. This continuous optimization in the field will provide for more knowledge on reducing emissions and a cleaner future for our planet. Business ecosystems are not new. What is new is that they are becoming increasingly data empowered. To realize complex opportunities, we need innovation beyond boundaries, democratized information and close collaboration between diverse players. Collaborative, data-empowered borderless innovation is how we embrace a world of exponential change. And that's what this podcast is about. Thanks for listening to Data Today, brought to you by Zühlke. I've been your host, Dan Klein. Discover more episodes of Data Today