Boeing acknowledges that production-quality issues meant the first 1,000 or so 787s delivered did not meet its specifications. But it insists the productions issues have been corrected, and that the in-service fleet is not at risk.
A Boeing engineer has come forward to challenge these claims, sparking a new round of scrutiny for the beleaguered manufacturer. Aviation Week Network editors break it all down.
Don't miss a single episode of the award-winning Check 6. Subscribe in Apple Podcasts, Spotify or wherever you listen to podcasts.
Discover all of our podcasts on our Apple Podcasts channel or at aviationweek.com/podcasts.
Transcript
Jens Flottau:
Welcome to this week's episode of Check 6. My name is Jens Flottau and I'm the executive editor for commercial aviation at the Aviation Week Network. We have talked about Boeing quite a number of times in the past few months on this program, but the company just does not seem to be able to get out of this loop of negative events. So we thought it's worth another deep dive and we're talking about the events of this last week, specifically the 787 program.
Just to recap, earlier this month a whistleblower, Sam Salehpour, claimed in a New York Times story that the way Boeing has been assembling the 787 is creating safety risks. Then, the Senate held two hearings on Boeing's safety culture in general and Salehpour's claims in particular. What senators and experts had to say about Boeing was not something any company would like to hear in public. We want to try and go into the 787 issues in particular today.
With me to discuss them here are Sean Broderick, our senior safety editor, who has been covering the Boeing crisis in much detail really since the MAX crashes in 2018 and 2019. And we have Graham Warwick on the program, our executive editor for technology, who can speak to potential challenges of building composite aircraft. Sean, let's start with you. You actually went to the Charleston plant to be briefed by Boeing and covered the hearings. What are your main takeaways?
Sean Broderick:
The main takeaways are that Sam Salehpour's whistleblower claims on top of everything that's happened to Boeing in the last... well, not only five years, but really in 2024 on the back of the Alaska 1282 door plug issue really got Boeing's attention. They quickly pulled together a briefing and brought some of us down in person to look at how they are putting the 787 together today and give us data on their outlook on the airplane's airworthiness. Salehpour's claims at their root are that manufacturing problems that are known now on the 787 and caused production to slow to a crawl in the 2021/'22 time period, and caused deliveries to stop for the better part of two years really, those manufacturing issues are being... The fixes that Boeing was supposed to put in place are being short-cutted and the assembly processes are still introducing risk on the production line.
Boeing has acknowledged that not only did the 787 assembly process have issues, but basically, every 787 out there was not built to their design specs. They didn't conform and getting them into conformance is they were able to get the FAA to sign off on their plan to begin deliveries again or restart deliveries in (what was it?) August of 2022.
The issues are, in the simplest terms possible, mainly gaps between pieces of structure in mostly fuselage barrels, but there are some other parts as well (door surrounds and things like that) where there were gaps. Tolerances are exceeded in the assembly process. There are a couple of ways that those gaps are normal in every aircraft assembly process and they are failed using shims usually. In Boeing's case, sometimes shims were not used when they should have been sometimes incorrect shims were used. And so the result, according to Salehpour, is that you have airplanes that long-term may have what we normally would consider fatigue risk, but in the composites world it's a little different than that, but basically long-term risk to the fuselage.
Salehpour also said that the process that Boeing uses to get pieces to fit together, a step known as fit-up force, basically means you're using force to get the fit between two large pieces of structure as tight as possible and minimize that gap. Boeing was using force that was outside of its own declared engineering specifications and, therefore, introducing potential strain into pieces before they were being fastened together. That's still with some gaps left in there potentially.
The takeaways are that Boeing acknowledges that every airplane out there in service has some of these non-conformances. It claims, A, that the airplanes coming off the line now and since August 2022 do not and it has the FAA's stamp on its inspection and prepare plan to justify it. But I think more importantly, long-term, that the airplanes in the fleet, while they do have non-conformances, there's going to be very little, if any, added inspections or modifications to do the work that is now being done on the airplanes before the inter-service. Basically, getting them into conformance is not a requirement to ensure the long-term... the 44,000 cycle life that Boeing has spec'd out for the airplanes. Even at high utilization, that's three decades of usage.
So those are the big takeaways. Boeing says the 787 fleet is safe, even though it wasn't built to conformance. They credit their tolerances, which were extremely conservative for this design, and the properties of composites for basically bailing them out, for the lack of a better term, on the in-service fleet.
Jens Flottau:
So Graham, why is it so difficult to put together these composite fuselage sections?
Graham Warwick:
I think it's probably best if... I'm not an expert on 787. Guy is. He's not... couldn't be with us today, but I think the way I would approach this is, why do we care about joins? So, we care about joins because we want the structural loads on the airframe. We want them to move freely around the airframe, so that they even out; they spread out and they even out and they don't introduce dangerous stress concentrations that could cause a damage over time or failure. Right?
So if you get a gust on a wingtip, you want that load that's induced at the wingtip to flow down the wing structure into the fuselage and get spread out over the whole fuselage. Ideally, you want a structure with no joins so that you've got those absolutely perfect load paths that are uninterrupted, but you can't build an airplane that way. Today, we can't build an airplane that way, so you have to have joins. Now as soon as you interrupt that load path, you interrupt that transfer of loads from one section to the next, so you have to try and come up with a way of joining the structure. That is the most efficient way of transferring the loads across the join.
Now, I'll get to why composites are challenges, but basically, a standard aluminum fuselage, a join is about 60 to 65% compared to then ideal single structure. If you bring a join in, it's about 65% efficient. You do get some concentration of loads, but composites by and large are only 40% efficient because they're bolted together at the moment and, therefore, the bolts are the sort of the key things that enable that transfer. But what you want to get that transfer of loads is to have the two surfaces of the two pieces of the structure to be in as close contact as possible. Ideally, you want an absolutely flat join that's so that, when you join them, the two flat surfaces completely butt up against each other. It's not possible to build something that accurately.
So let's go and look at why composites and metals are different. It's a couple of key reasons. Aluminum is bendable, so if it's an old style sheet metal airframe... and we've done this for decades and decades. When you put two pieces together and if you don't quite fit, you just bend the metal a bit. You're not talking about huge bends, but you're talking about just massaging the metal a bit till you get a good fit. And because it's fairly ductile, that doesn't really increase the stress where you've bent the metal because aluminum doesn't care where the load comes from. It could come from side to side, up and down, through, across, whatever. It doesn't care.
Composites wants the loads to go in in the direction of the fiber. They do not like loads that go through the laminate or across the laminate where there are not fibers that could carry the load. So you have to design them carefully, the loads. The loads and the material are the same. The second big way of making metal airframes is you machine them. So if you machine them, you can accurately control all the adjoining faces, everything like that, because machining it so you can have a much tighter tolerance on it.
Now with composites, the biggest thing with composite with metals, you make the material, then you make the part with composites, you make the material and the part at the same time. So when you make a piece of composite section, you put it on a tool, you bag it up, you put it in an autoclave, and you cook it at high pressure and the plastic resin melts. It binds the fibers together and you get a single piece of structure out the other side, but only one side of that, that part, has been on a hard tool. The other side is up against a bag or something like that, so it's harder to control the variations of the thickness of that part. You're going to get variations of thickness because it's not a precise process. You do not have a controlled flat surface on both sides of the part.
Secondly, it's very stiff, so you can't bend it when you get it onto the assembly. Not only that, there's a thing called springback. It doesn't matter how hard you try, if you have a curve in composites, when you take it out the tool, it will slightly unbend. It's called springback. So you've got to try and design that into the part so that when it springs back, it springs back to the correct shape, but it's very difficult. So that's why Sean talks about conservative tolerances. What you do is you look at all the pieces joined together and you add tolerance buildup, and you say, "I can only make that part within this tolerance. I'm joining it to that part that's only made to this tolerance. I'm joining it to that part that's only made bent to this [inaudible 00:11:15]." It all gangs up to where you get to a major joint, you've got all this stack up of tolerances, and you've got to put two parts together.
These are bolted joints, so the issue here is that when you put the two parts together dry, you literally just push them together, you'll see gaps. So the next thing you do is you apply what they call fit-up forces, which is like a temporary fastener. You put a temporary fastener in and you tighten it to a very, very strictly controlled bounce pressure, and then you're supposed to measure the gap that remains after you've put those fit up forces on. You measure the gap that remains. That is the gap that in essence won't close. That's where you put the shim in.
You size the shim to fill that gap. You put the shim in that. You unbolt everything. You put the shim in. You bolt it back together again, this time with pull-up forces, which is the final force that will hold the airplane together with permanent fasteners, and then you tighten it to that pull-up force, which is the final force. And then, you theoretically have two surfaces that are in contact all the way around the joint, either directly part-to-part or through the shim. They're in contact. Loads can pass through and you've got a safe structure.
So what Salehpour was saying was that, in essence, they were applying too high what they call the fit-up force, which is the initial force when you're trying to work out, what is the gap? So you apply that fit-up force with temporary fasteners, you measure the gap, you make the shim, and then you put this thing back together again and you apply the final pull-up force. He says that they were pulling the structure far too much before they measured for the shims and that was against what Boeing's own design rules were. He says that that excessive pulling-up before you measure for the shims has this risk of introducing stress concentrations that could cause damage over the life of the part.
And as Sean again said, it's different with composites. In composites, it's about microcracking of the plastic resin. It's about pull-through and other types of failures around the fasteners. It could even be about the fastener itself just comes under stress and it will break over time. It's not the same as a crack forming as a whole and spreading through the aluminum frame. It's different. So he was saying that that's where the problem was. That's where the excessive forces were introduced and that was because Boeing was not following its own design rules.
Jens Flottau:
And does he claim that they did this to minimize the size of these gaps or was it just a mistake?
Graham Warwick:
Maybe Sean can answer that. I would have to say that this is... We are still in the infancy of composites. This is a fuselage barrel, which nobody's ever done a fuselage barrel before, and you've got a workforce that, if they have any experience with aviation, experience with metal airframes, not composite airframes...
He talks at some point about people jumping up and down on top of things to make them fit. I don't know, but it sounds like a lack of experience in the manufacturing side certainly may have led them to do things that they were not supposed to do. But I didn't hear any implications that the design side of Boeing didn't write the specifications correctly. It was all about that the specifications were not followed in the manufacturing process.
Sean Broderick:
Either were not or could not be.
Graham Warwick:
Well, that's another possibility, yeah. Can I just jump in here? Because I just want to give you my reaction to Salehpour. I thought Salehpour was a very credible witness, clearly an engineer. He came across as an engineer, but very particularly, he's a quality engineer. His responsibility is to make sure the aircraft is manufactured the way it is designed.
When he talks about risks, he talks about risks introduced because the aircraft was not manufactured the way it's designed. When I listened to what Boeing is saying, they're actually addressing the design. They're saying, "Yeah, we didn't make it the way we meant to, but look at the in-service experience. The design is robust enough that it's not causing problems even though we didn't make it well enough." Now, that's not really a great excuse, but they're actually talking about two different things. He's talking about risks from manufacturing. They're talking about robustness of the design when it isn't built right.
Jens Flottau:
And they point to the maintenance experience that they've gained so far over the 12-year life of the in-service fleet that didn't show any reasons for concern. Right, Sean?
Sean Broderick:
With the composite fuselage structure, yes, and they're very careful to point to results from... I think it's that, "We've had eight heavy checks, which come at about 12 years." They've had ten supplemental structural inspection processes on airplanes in service to help gather some of this data. They also point to their fatigue test done over five years in the beginning of the airplane service life: 165,000 cycles. Again, that's four times the design service life, which is longer than any of them is going to fly anyway.
They point to all of that and say, "All of the data from these shows us we had some issues on some metal parts or maybe some metal-to-composite combinations, but nothing on the composite fuselage that's the subject here." And as Graham talked about, we're talking about one-piece barrels. Most of the issues around these one-piece barrels, and where Graham was also talking about the manufacturing processes barrels, the side with the hard tooling versus the side with the balloon or the bag, that's one of the primary issues that Boeing is still working on in manufacturing. They actually have to hand sand the edges of these barrels along where the connection is where the joint is going to be to get the smoothness on one side, or the smoothness on both sides into their tolerances, which is no more than a variance of five thousandths of an inch over a five-inch span.
The Boeing down in Charleston, the Boeing engineers have talked to us, said sometimes getting these things right even today can take days. And so, while Boeing has identified... It says it's identified all the manufacturing non-conformances is correcting them, the next question is, how on earth they scale this up to the kind of rates that they want to produce the 787 at? And then, what about the next airplane if they decide they're going to go with a similar design? Those are questions for another podcast, but they're some of the ones that came to mind when we were down there in Charleston watching this process.
Graham Warwick:
Can I just interject here? One of the things, it's absolutely about the way you make these composite components. If you go to the F-35, which has got large composite structures, they actually have to... because of the tight tolerance on the outer mold line, the outer surface of the aircraft, they have to machine every single joining surface.
So literally, you go through the process of making the composite part and then it goes into a milling machine and they have to machine the... See, the only way that you could guarantee the tolerance that would allow almost a flawless joint on a fuselage barrel is to machine every single join across, and when you're talking about... Well, I don't know production rates. If you're talking about a narrowbody (what is it?) 40, 50, whatever, a month or... I don't know what it is. It's not that I'm not a commercial guy, but it's just untenable to think of doing that. So you have to come back to that's why they use shimming. They didn't want to use shimming. I suspect there were some starry-eyed guys in the design office that never thought that shims would ever be a thing, but this is the first generation. And as Sean says, we're going to have to do better next time and it is as much about how we manufacture the next one as it is about how we design it.
Jens Flottau:
So speaking about improving, the aircraft that come off the line today, are they within the specifications that Boeing has put in place initially, or are they still out of tolerance?
Sean Broderick:
Well, it's an interesting answer to that question. I think it's safe to say they are within Boeing's current design specifications. They do conform or else they couldn't be delivered. Now, what has happened in the last few years is, as Boeing has gone through what it calls its tip-to-tail join verification process, which is its catch-all name for this, it has changed (with FAA approval and data to back it) some of the specifications on gap sizes. So the five thousandths of an inch that gets thrown out as the general tolerance level, some of those have changed based on Boeing's data.
And again, that goes back to composites and the robustness of composites. So all of the airplanes being delivered today and being manufactured since around the late summer 2022 do conform, but there are still about, I don't know, 35 to 40 airplanes that built up before those changes were put in place and before the FAA approved Boeing's plan that are going through the rework process out in Everett. That rework process includes some of the stuff we talked about, but at one join, at the 47/48 join, which is at the back of the airplane, they have to remove every single fastener from the join and do what they call a through-hole inspection with an angled feeler gauge to see what the gap is adjacent to each fastener hole along the composite strap that holds the two pieces of fuselages together. It's over 2,000 fasteners. It's an incredibly time-consuming process and it helps explain why so long to get those 120 airplanes that built up by the end of summer 2022 into operators' hands.
Graham Warwick:
That illustrates what I was talking about: that Salehpour is a quality engineer and a quality engineer's job is to make sure the aircraft is manufactured as designed. If Boeing takes its experience from 300-and-something 787 and changes the design specification to make it easier to build, that's a different issue. You would just have to assume that the design engineers would've to go through a safety analysis to say, "We can relax that requirement because we have a greater safety margin than we thought in the original design."
That kind of thinking happens to all airplanes all the way through their life. Airplanes get heavier through their life. They'll release a gross weight increase. That gross weight increase is because when they designed the airplane, they were conservatives. They put in margins that they found with service experience they didn't need. It could be argued that Boeing is taking its manufacturing experience with 787 and changing the design to relax the requirements to make it easier to manufacture, but as long as it's based on data that shows that that does not induce safety issues or reduce the long-term safety. That's something that we do throughout the life of airplanes in different ways, but it's something that we do to airplanes. We change as we learn what the real data, that we change the margins to either get more performance or take cost out, or take time out, or do something like that.
Jens Flottau:
But doesn't Salehpour also claim that a change in the way that the aircraft's been built, put together, has introduced that risk?
Graham Warwick:
He said, "We've built 300 of these, we should be in a stable situation." You get to 300 airplanes and you do not have stable processes, stable repeatable standard processes, then you've messed up. And he said that the process... he implied and he used the word unstable. I think he was referring to the processes on the 787 being unstable in its current state. He'd just saying, "After 300 airplanes, we shouldn't be in that condition." So stable repeatable processes are something that you drive for in aviation.
Sean Broderick:
And Boeing didn't help matters by not only relocating and consolidating all the 78 production into South Carolina, but they also made, again, for long-term benefits, substantial changes to the 9 and the 10 that required... most of them in the ass section, that 47/48 section, that were so extensive it required an entirely new fatigue test on that section, another 165,000 cycle test on the part of the airplane. It wasn't the whole airplane, but on the back of the airplane.
So the airplane has evolved and Boeing has changed the production location. These airplanes, the production line hasn't been open for that long in the grand scheme of things: 15/16 years. So they've changed a lot of things in that time on a design that's... They were first out with the full composite airliner. Airbus with the A350 has taken a similar approach, but they took a... Their design is more conservative using panels instead of one-piece barrels, for example. So a lot of lessons learned, I think, in this process.
The last thing I'll say from Boeing's standpoint, they can't afford to be wrong. They can't afford for this engineer's concerns to be right. Engineering concerns, I think internally are nothing new and I think they happen far more than we understand, but the engineers, they're not always crying wolf. There were engineers early on in the 787 battery development process that had concerns. They proved to be correct. There were concerns when the MAX was being designed about the logic of having two angle of attack centers that were only two and that didn't work in tandem, and that proved to be prescient. So, Boeing can't afford to be wrong on this with their claims that the long-term safety of the 787 fleet is not a risk.
Jens Flottau:
And Salehpour's presentation at the hearings was quite dramatic in some regards. He was saying, "The fact that I've come out with these concerns will save many lives in the future because it'll force changes." What do you make of that?
Sean Broderick:
He's adamant. I am waiting to see. He sent a letter to the FAA, well, through his law firm, through official whistleblower process in early January, and it cited some documentation that has yet to be made... that we have yet to see. I'm curious to see some of the documentation, including the data on a subset of airplanes, that he has and maybe some exchanges internally.
He talked about a white paper that was written by some fellow engineers that have similar concerns, so I'd like to see some of that before I make further judgment on what he said. It's very difficult to... I'm certainly not going to call him uncredible, but all we have now is he's making claims. At least at this point, at least Boeing dragged us all down to... or invited us all down to Charleston to show us some things. Maybe they're just better at PR than Salehpour is or maybe they're just better or good at PR, period, but at least they stepped up. Aviation Week had been asking for demonstrations on what these gaps are and how you check for them and how you correct them for two years. So if there's anything I can thank Mr. Salehpour for, it's for getting Boeing to finally get us down to Charleston to see some of that.
Jens Flottau:
We have to wrap up quickly, but I have one final question, which is, what are the next steps? We went through the hearings. We've got his claims, Boeing's rebuttal, what's next?
Sean Broderick:
The FAA will certainly look at what he has brought forward, but again, the FAA is in a tough spot because they approved everything that Boeing... They really put Boeing through the wringer when it came to gathering data and then approving the inspection and fixed process for the join verification. The FAA has seen everything that Boeing is doing, unless Boeing is hiding something, which Salehpour said is not out of the realm of possibility.
So I think the next steps are to look more closely at what Salehpour has and try to look at it with the same skeptical eye that we can now look at the 737 MAX development exchanges, all of which came out after two fatal accidents, which underscored that certainly something went wrong back there. We don't know if anything has gone wrong on the 78. We have accusations that it has. So I think it's going to be very careful scrutiny both from the FAA and from anybody else who can get their hands on any of this information to see if any of it holds water. We'll be watching 787 fleet closely and I'm sure the airlines will be looking at their inspection results very closely for every airplane that's sitting in a hangar.
Graham Warwick:
And the hearing, the congressman did make it clear they're going to have more hearings, so Boeing is going to be in the public eye a lot more.
Jens Flottau:
In general, and I'm sure also here on Check 6, but that's all we've got time for today. Special thanks to our podcast producer in London, Guy Ferneyhough. Special thanks of course to Graham and Sean. Don't miss the next episode by subscribing to Check 6 in your podcast app of choice. Again, thanks for listening and goodbye.
Related Content
