Lilja Scheel Birgisdóttir, Reliability Engineer at Icelandair, talks with Sarah about the immense complexities and interdependencies of airline operations and the role predictive maintenance will play in the airline’s future.

Sarah: Welcome to the Future of Field Service podcast. I'm your host, Sarah Nicastro. Today, we're going to be talking with Icelandair about their exploration of predictive maintenance. This is a conversation that I think you'll find quite interesting. Because when you talk about service complexity, I don't know that there is a more complex environment than the airline industry. I'm excited to welcome Lilja, from Icelandair, to talk with us today. And Lilja, I'm going to let you introduce yourself fully, because I know that I would not properly pronounce your last name. So, please tell everyone about yourself.

Lilja: So hello. I'm Lilja Scheel Birgisdóttir. So, we are from the airline, Icelandair. I have my team with me. We are a team of reliability engineers. We consist of three. And it's Harpa Rún Garðarsdóttir and Unnar Már Sveinbjarnarson. And then, we have our IT specialist with us who is Bjarki Elíasson. So, if I just introduce Icelandair briefly, it's an airline that has celebrated over 80 years, birthday, if you can say that. So, we track our history back to 1937. So it's only 34 years after the Wright brothers took off, we took off, not in the form that we are today. It has a longer history of mergers, but we've been flying under the name of Icelandair since 1973. So, not many airlines that can say that, especially when we're just a tiny airline flying from a tiny island in the middle of the ocean.

Lilja: We have a specific business model which has been very successful. It's called the Via. I'm not sure on the pronunciation, but Via market because we fly from Europe to the US, and we offer the great deal of free stopover in Iceland for up to seven days, which is kind of intriguing for many people, and at no extra cost. So this is a unique model that not many can replicate based on our geological position. What is kind of unique about us, I would say, but maybe the global market does not know, is that we have what could be explained as the heart of Iceland there, or heart of Iceland. And we have felt it very strongly, especially in the 2010 and '11 volcanic eruptions, from the very well-known volcano, Eyjafjallajökull, which every reporter struggled with pronouncing the name of back then.

Lilja: But throughout this Corona situation, we have felt it, again, beating strongly when our day-to-day work changed a lot overnight. And we have had to say goodbye to over 2000 fellow colleagues. So, it's been a hard time. But again, we have been feeling this strong unity. And amazingly, Icelandair is so well liked, and by the Icelandic nation as well. So many of them, they experienced their home when they step on board of our aircraft on their way home. So, they feel the Iceland spirit and the Icelandair spirit. So, that's a little bit about our company, who we are. If you want to know, understand, what we do as reliability engineers? So, what we do is collect and monitor technical data, so that we can evaluate how well our planes are doing. We monitor technical dispatch reliability, or/and... So, we're monitoring basically the health of our fleet. And if we are seeing any off trends, we notify someone who can take action on those. So, that's a little bit about us.

Sarah: Excellent. Well, thank you, Lilja. I have heard of the via structure. And it is very compelling. I have not had an opportunity to visit Iceland, but would love to. So, once travel picks back up, I will put that on my list. And I know that-

Lilja: You should.

Sarah: I know that for everyone in the airline industry, 2020 has been a very, very difficult year. So, certainly appreciate that. And I think most of the conversation we'll have today is not necessarily talking about business as it's been recently, which is uniquely difficult in a lot of industries really, but talking about the complexities in the airline operation and for Iceland in a more normal state of business, which I think is what most folks intuitively think about when they think of flying, because a lot of people are used to also being able to travel far more than we have over the last year.

Sarah: Well, thank you for the background. So, what I do want to talk about next is some of that complexity. So, we had a call to kind of set the stage for this recording. And I had shared with you all that I've never really thought about the weight of complexity on an airline. I guess I should have. But I just really have only thought about it through the lens of a passenger or a consumer, and not necessarily all of the different aspects of complexity. Some, I think, are intuitive because they're more of the customer facing things that you would notice or experience. But there's far more to it. So, I want to just talk about some of the different facets of complexity that makes managing and optimizing an airline operation quite, quite challenging. So, to start, let's talk a little bit about... The first area is customer expectations, right? So, maybe talk about that one first?

Lilja: I mean, as a customer, you want good service. You want reliability. You know, you want to be there alive and on time, basically. I mean... and smooth and enjoyable travel experience.

Sarah: Yes.

Lilja: And that's where on-time performance is a critical thing. But, I mean, it's the same for every airline. It's the same expectation. We have the same customer group, basically. So, I mean, that's the main thing.

Sarah: And it's the most intuitive and easy to understand, right?

Lilja: Yes.

Sarah: Because the vast majority of us have been on a flight that ends up delayed or canceled. And then, you... So, that frustration of the interruption to your life, and your schedule, and all of that is... It's the easiest to comprehend in terms of the complexity. But let's talk about... Then, when there is a disruption, obviously it impacts the customer experience. But talk a little bit about what happens... It's really a chain reaction, so there's a huge trickle-down effect of one issue throughout the system. So, what are some of the things that happen or areas of impact once you have some sort of issue, or failure, or downtime?

Lilja: I mean, what's happening is there are a few cases that can take place. I mean, you can have one part that is faulty, and the... In the cockpit, they see a light, and they just... They know they can't take off. It can be that it's late from another flight. It can be that it's late from a maintenance. So there are few things that can happen that causes a delay, or an AOG situation, aircraft on ground. I mean, in the case of we have a faulty part and they have the light than the cockpit, we're seeing 15 minute delay, and up to a few hours. This can lead to an AOG, where your flight is delayed, canceled, whatever. And sometimes, we can resolve it right on time, especially if the aircraft is in their home base. It's easier. If the situation occurs at an out base, it's a bit more cumbersome.

Lilja: And, I mean, for the customer, it can lead to missed flight, missed train. They have to book a new hotel. There's so many things that can... the trickle-down effect. So, I mean, if you have an AOG situation that's a longer time. The aircraft cannot take off. It can... You might end up where you have to get a new aircraft to pick up the passengers. So, then that's lost revenue when you have to fly a new aircraft empty to pick up all the passengers, and then fly the broken aircraft when it's been repaired. So, I mean, this is... There are many things we need to consider when this happens.

Sarah: Right. And the other thing is, if you think about the kind of interdependencies of air travel, right? Like you can't... When you have a delay, or you have an AOG situation... Now, I sound like I know what I'm talking about. Then, you can't just, "Oh, okay. Well, we have it fixed. So, let's just take off." Right? I mean, you're reliant upon all of the other travel happening in the air, all of the other... the air traffic. You're relying on a lot of different things. And so, to your point, there's an issue when it comes to customer satisfaction. So, how frustrated are your customers getting in those situations? But there's also a very real issue of cost complexity. Right? So, to your point, I've been in situations before, if it's extreme enough, where a flight's canceled or what have you, you end up getting some sort of compensation from the airline. So, you have that cost.

Sarah: You have the cost of flying empty planes, and all of those things. So, it becomes quite intense, in terms of the impact of that. And I think you also can't under emphasize the criticality of the number one objective, which is keeping everyone safe. Right?

Lilja: Exactly.

Sarah: So, I always think about... I've been in... Anyone that's been in a situation where a flight is significantly delayed or canceled understands how frustrating it is. But I always find myself annoyed with people that throw a temper tantrum about it. And yes, it's frustrating. But the number one thing is you don't want someone to take off in an airplane that they know they shouldn't, or something isn't properly taken care of.

Sarah: So, it's always like... Yes, I know it's frustrating, but they're acting in our best interest, even though it's inconvenient. Right? So... Anyway. But the other thing is... that I had never really thought about prior to our conversation is some of the complexity that exists in being able to rectify issues, or make necessary repairs, to be able to move along. And so, talk a little bit about... I guess I just never thought about the fact that, if you have an aircraft in some location, you can't just have any mechanic come and do a fix. You don't know if you'll have the part you need. So, talk about some of the complexities when it comes to the regulatory side, and some of the inventory, and those sorts of things.

Lilja: Yeah. I mean, we have a lot of regulatory bodies. We have FAA. We have EASE. And they have different regulations. And Icelandair, flying in both regulatory bodies, we need to abide to both. So, then we of course have the... from Boeing, everything we have to just obliged from there. So, everything is highly regulated. And you wouldn't believe the paperwork that goes with one aircraft. I mean, it's tons of paper.

Lilja: So, everything that's done has to be written down, signed off. And the person that's signing it off, as you said, has to have the permission to do so. So, the person has to be trained and licensed to do this thing. Even though you're an aircraft mechanic, you're not necessarily have the license to work on avionics, so the computers on board. So, if something comes up, you need to be sure that the person has the specific certification before they can come aboard and do the work. So...

Lilja: And the parts that we have, even though we have two 757s, Boeing, standing side by side, doesn't necessarily mean that we can use the same parts in both, because they have different specs, they have different modifications. So, they are... We might need two different parts, even though the same one failed. And these parts aren't just laying around in every stock room. We might have a AOG situation in Boston, but the part is available in Europe. And you have to transfer the parts. And you need to make sure that the part is certified properly, with the paper work that we require, with the mud status that this aircraft requires. So, there's so many things that need to align when something comes up.

Sarah: Yes. Ooh, it's not a job... I don't think I would want... I don't think I would want it. It's a lot of pressure. So obviously, you have to handle those situations when they arise. But the more you can avoid them through both traditional maintenance and what we'll talk about in a bit, predictive maintenance, the better off we are. So, the ultimate goal is to minimize and eliminate as much as possible, any sort of issues.

Lilja: But I must add that. I mean, it takes a lot of people to cover these things. So, it's not only one man job, thankfully.

Sarah: Yes, yes.

Lilja: As you say, you would not want this job. You're only getting a tiny part of the whole scenario.

Sarah: Right. Right. I don't... I just don't... I don't know that I would want... I just don't know that the airline thing would be for me. I don't know. I get anxious as a flyer, let alone being responsible for all of that. But it isn't... It is... over... holistically, not just yourself, but everyone in the industry. It's an important job.

Lilja: It is.

Sarah: Like I said, I mean, it's... People's lives are in your hands every day. And when you look at some of the industries where... You talk about mission critical situations. And sometimes, people refer to that term in terms of downtime costs money. And yes, that's important. But when there are situations like with this, or in certain medical applications and things like that, where it's lives, number one, I mean, you're talking just about a different level of importance of, of everything working, and of managing and optimizing, and paying close attention, and, and all of that. So, a lot of respect for what you're doing.

Lilja: Thank you.

Sarah: So, let's talk a little bit about the kind of historical and present-day maintenance world for Iceland. So, you use IFS Maintenix to manage the planning and orchestration of all of the maintenance of the aircraft. So, tell us a little bit about how that works, and the value that IFS has contributed to Icelandair's operations.

Lilja: Yeah. So, we've been using Maintenix from IFS since 2014. We started the implementation in 2013, so... Wow. Eight years almost? That's crazy. Yeah. We use Maintenix to keep the aircraft everywhere. They keep track of all our scheduled maintenance activities, planning actions, and... Yeah. So, it's covering most of the processes that we do.

Sarah: So, you have Maintenix in place to manage the maintenance operations. But you're obviously continually looking for ways to minimize and eliminate any sort of delays and AOG situations. So, this is sort of the thought process behind investigating predictive maintenance for critical parts on the aircraft. So, tell us a little bit about how you view predictive maintenance potentially helping Icelandair.

Lilja: So, if we get a tool that can predict maintenance, we can improve our whole planning overview. We can start ordering parts beforehand, so cost saving. We can plan the maintenance action before it happens, before we get an technical delay or AOG. So again, we're saving costs and we're increasing our on-time performance. And, I mean, we could start... We could send the aircraft with the parts that we know is about to fail, so that it can be replaced wherever the aircraft is. So, I mean, we could take... I'm not saying that, that's an ideal thing to do. But we could know beforehand what we can do, when we can do it, and how we should do it. So, it would give us so much more insight and preparation time than we have today. So, in the long run, we would get shorter technical delays, reduce AOG situations, and reduced problems when we need to order parts that possibly is in somewhere in Europe, or America, when we need it in Iceland. So, there are so many things that... predicting the need before it happens.

Sarah: Okay. Tell me how this works on... You have an older Boeing fleet. Right? So, how do you determine what components you want to use predictive analytics with? How do you sort of put that in motion with the fleet that you have?

Lilja: Let's say there are certain part groups that we know that we don't need to monitor, because they're just you use and replace, so fully excluding those. So, we are looking at multiple part groups and part types. In terms of an older fleet... I mean, we know that our 757s are kind of old. The bad thing there is that they don't have as many sensors as the types that are coming out today. So definitely, that does not help us in the predictability. But with the coming fleets, we have a few max, 737 max. So, they have more sensors. So, we know we're going to get more data there. But then again, it's also just mathematics and statistics. So, we know that the models that they make in such a system are using historical trends.

Lilja: So, that helps us. Even though the 757s do not have sensors to assist with the predicting, we know that we have statistical data with us. So, that's going to help. And it's not impossible. So definitely, there is some future there. And it's so much fun to be in this place because, all of a sudden, we're taking part of the future happening.

Sarah: Yes.

Lilja: When we see... We're taking the step into predictive maintenance. It hasn't been as much viability before as it is today and tomorrow.

Sarah: And why do you think that is? Why do you think predictive maintenance is more viable today and tomorrow than it has been for Iceland air historically?

Lilja: I mean, the technical advancement has just been so great in the past years. Computers are getting more powerful. I think also, because of how highly airlines and maintenance is regulated, it's not been so much in the computers. I mean, every work that we do... Some airlines do have what's called e-signature, where the sign off on the papers are online, if you can say that. We do not. We have not implemented that yet. So, I mean, getting the data into the computers is just happening today. And it's really slow process because it has to fulfill all those rules, regulations. We have the CAAs that have to approve everything. So, it's a slow process.

Sarah: You wouldn't want to be an early adopter in the situation you're in. You need something that is more proven, and has... I can understand that. So...

Lilja: At least you have to have a lot of money if you want to be the first one.

Sarah: Yeah. That makes sense. Okay. So, you have a proof of concept around predictive maintenance in place with IFS to explore this more. Tell us about the project and its intended objectives.

Lilja: So, the basic goal is just to build a view which displays some fancy graphs and information tables. So, that's the thing. It's going to be... As we've stated, it's going to be so useful for us to be able to see the predictions down to the serial number of the components. If that part is expected to fail within a certain time, this information would then be fed to our planning team or the maintenance control for further action. This seems very simple when I say it like this. But the calculations behind the predictability, the model training, is... Those things are fairly complicated to do. And it has to be good data that is fed in. So, that's also a very critical factor. So... But the goal for us is to be able to predict a component failing before its time.

Sarah: Okay, you touched on this a little bit, but I want to dig into some of the ways that this really benefits Icelandair and its customers if it works well. So, if you are able to successfully migrate to this predictive model, it... To me, it gives you better control, and it gives you more time to react, and it takes a lot of the uncertainty out of the situation. Right? So rather than, you have a flight land in Boston and someone calls and says, "Hey, uh-oh. This light is on. This part, it has an issue. And that part is in Iceland, or that part is in Europe. And you don't have the right person there to fix it. And you need to..." You know? So, there's all this time then that passes to align the right situation to fix it, and then comply with the air traffic and everything to get the flight back in the air.

Sarah: So, if you knew, to your point, "Okay. Well, this part is nearing end of life," or "We know that there's going to be an issue," you can orchestrate a repair before that failure occurs, or at least put yourself in a position where you have the appropriate resources where they need to be. So, you're more able to precisely align resources when, or even before, failures or issues occur. So, talk a little bit about that. So, you can... The different components you're... And you're focusing on tracking components that are more of those critical components, so things that would ground a plane, not the things that you have, like you said... I can't remember the term you used, but like a use and replace, that are more readily available. Right? You're talking about the things that would take time to fix. So, with the predictive model, you would have the visibility into when and where things will be occurring so that you can align the resources to those situations. Right?

Lilja: If something comes up for parts that are flight critical, you cannot fly. But if it's not a critical part, you can... I'm not completely familiar with the process, but you can ask for an extension of life, so that... I mean, you can get permission to wait for two weeks or fly the airplane home. So, there are processes there. But these are the components that will ground the aircraft. And knowing beforehand if these components are going to fail... Yeah, we're going to reduce costs. We're going to reduce problems and unpredictability. It's going to be great with failure prediction

Sarah: And you're also obviously improving the customer experience. Right? I mean, anytime you're able to minimize or eliminate a delay or a grounded flight, you're improving that as well. Talk a little bit about how the data could be leveraged over time to analyze patterns that could help Icelandair, so looking at patterns of faults and failures and how you could use that information within the organization to make changes.

Lilja: That's basically the goal of the reliability engineer. That's to identify trends and patterns, and try to see why they're happening. And we have always, what's called a reliability meeting where there's a reliability control board that meets. And we inform them of everything that's happening, trends, and things that we are seeing, the things that we have seen. And then, this boards, which persists of... I'm not very good with stature... status names, but like all the people that are required. We have pilots. We have mechanics. And everyone that's needs to know about the things and have something to do with it, they come there. So, it's a joint board where we can discuss these things. They can decide on future steps, or like the board can decide on future steps what needs to be done.

Sarah: So, over time, you can use the data that you're gathering in the predictive maintenance program. And if I'm understanding correctly, it also... The more data it gathers, the more it learns. Right? And the more accurately it can predict.

Lilja: Yeah. That's how the predictive maintenance system will work. It's a learning process.

Sarah: So, you can use this to look at different ways to make changes within the business to, again, just kind of operate more effectively because you have that far improved visibility. Right?

Lilja: Yep.

Sarah: Now, what about external implications? Once you have data that maybe shows you things that they... Let's say you... I'm just making this up. But let's say you have some part that you're noticing is continually failing prematurely, or you notice that there's an issue. Are you also able to take that data and leverage it with your external partners to make improvements out... kind of outside of the business?

Lilja: Yeah. When we see such a trend, we often look at which repair shop has been servicing the parts. And if it's just this part, sometimes we just have an odd one out. And then we just take it out of operation. We don't want to see it anymore because it's unreliable. It's costing us a lot. So, if it's just a single part, we can just throw it out. We just scrap it. But sometimes, we see it with the same part number. They're keep failing again, and again, and again, before it should. Then, often we look at the vendors to see if maybe they are not doing their work properly. Is it always... Are they all coming from the same vendor, or multiple vendors? So yes, this information will help us also to identify those. And this is a lot of costs that's... Parts cost a lot. A repair costs a lot. Everything in the airline industry costs a lot, except for the airline tickets.

Sarah: Right. So, there's significant opportunity to optimize and save. And it's interesting. There's a lot of potential value here in terms of you have sort of the customer facing benefits. You have the benefits for Icelandair, in terms of more visibility, better ability to align resources, cost savings, and all of those things. And then you have the ability to leverage the data externally, where it's relevant to improve some of the vendor and partner relationships, and give them feedback as well if you're able to pick up on those trends. It's really interesting Lilja. And I think that... I understand your point about why, in the airline industry and for Iceland in particular, you want to move pragmatically. You want to take your time and do this the right way. But our platform talks across industries. And this move to predictive is a huge, huge trend.

Sarah: And I think, to your point, in the last few years, the technology has just become so much more accurate and more accessible to people for a variety of different reasons. And it's exciting to see it sort of come to fruition in different industries, in different use cases, because it is really powerful. And the way that it will change how Iceland operates and how different businesses operate is really exciting. So, I appreciate you coming and sharing. And I'd love to have you back in a bit, when you are further along in the proof of concept, and talk about some of the things you've found, and some of the ways you're using the data and putting it to work, and kind of talk about the progress. I think that would be really cool.

Lilja: Yeah, definitely.

Sarah: Well, thank you so much for your time. I appreciate it. And thank you for sharing.

Lilja: Thank you.

Sarah: You can learn more about predictive capabilities and other trends by visiting us at futureoffieldservice.com. You can also visit us on LinkedIn, as well as Twitter, @TheFutureOfFS. The future of field service podcast is published in partnership with IFS. You can learn more about IFS service management by visiting us at www.ifs.com. As always, thanks for listening.