I closely watched these threads and found following consent:
The Barrel approach (B787) is generally considered superior versus the Panel approach (A350).
Among the many advantages following are mentioned often:
Cheaper and fully automated production
High integration from the beginning
Lighter (because of the following two points)
Less fasteners
Fewer joints
In this thread I want to discuss why Airbus is on the Panel track. I will support this by substantial sources and ideas. Among the most heard reasons why Airbus plans to use Panels we often heard so far:
Airbus has not the composite know how
Airbus can not use Barrels because of intellectual property issues
Airbus needs to go the fast approach therefore has not the time to make it 'right'
Airbus has not the logistics to transport Barrels
These all imply that Airbus needs to choose a technologically inferior solution because of surrounding requirements. I don't want to add much comment on these but want to raise the question: could there also be technological advantages that we just don't know?
I just mention in short the possible advantages that were listed so far:
Easier logistics when transporting panels instead of barrels
Longer panels than barrels possible
When seeking for other advantages quite often there were gentle hints by WingedMigrator but usualy he only earned ingorance or even opposition ( Airbus A350 Cfrp Panel Construction (by WingedMigrator Jun 22 2007 in Tech Ops) ). First I want to share what I gathered around of what he suggested. Let's develop a line of reasoning:
With Panels the outside of the fuselage lay in the Mandrel. Therefore you will get the smother surface on the aerodynamically important side of the wall (or at least better control on the outcome during manufacturing). Kaneporta1 once raised this point.
With barrels on the inside you can only vary the thickness to some degree and have limited possibilities to built 3D structures (for the B787 only stringers)
Having the outside in the form you are free to control the structure complexity on the inside and built 3D structures.
Thus you could have the idea to integrate and co cure various inner construction elements from the beginning. This could be:
Stringers
Frames
Window Frames
Door frames
Reinforcements that can not be built only by thickness change
Crash absorbing construction elements
I made a graphic that shows a panel with fully integrated stringers and frames:
It would come as one piece out of the oven. The number of fasteners is greatly reduced. The splice plate on one side could be made from the panel itself thus eliminating further rows of fasteners.
The built process to produce such panels would be fully automated. For final assembly the four panels just need to be connected with long rows of fasteners, and that's it (moderate handwork to get a complete fuselage barrel made out of 4 parts). In contrast with the barrel approach you get little more than the skin and you need to put in a large number of frames (more handwork to get a complete fuselage barrel consisting of many more than 4 parts).
The next picture shows a segment of a fuselage with one barrel joint. Red dots are fasteners.
Picture 1: Barrel approach
Picture 2: Panel approach light (only frames and stringers co cured with the panel)
Picture 3: Panel approach max (Stringers, frames, doorframes, window-frames, reinforcements all co cured with the panel)
Conclusion:
In the light of these ideas the picture changes dramatically. From the listed advantages for the barrel approach (easier production, Lighter, Less fasteners, Fewer joints) each one can be disposed. The panel approach leads to higher integrated parts allowing cheaper (more automated) production and final assembly. There are fewer fasteners and joints (if you count the joint between each frame and the barrel too).
Availability:
You see the panel has the clear potential to be technically superior. The question remains whether today composite technology is mature enough to produce such a thing (how complex can structures be built up on the panel inside?). If one day the industry succeeds to realize the "Panel approach max" as outlined above you will for sure see Boeing change to the panel approach!
Now how much of this technology is at hand for Airbus to produce such structures?
So far in this post I summarized what A-Net has brought up so far and what the advantages of panels could be. In the second part of the post I want to share all the sources I have found around this topic. A few google-searches yield a lot of interesting stuff regarding this question. I have found some treasuries of links that shed a lot of light on these questions. A general review of public information available on the internet allows following statements:
- There is a huge amount of know how for composite fuselages available in Europe
- Also crash worthiness is covered well with studies
The first link reveals about a crash worthiness study made in Holland in 2001: http://nlr.nl/smartsite.dws?id=4366
Very similarly to the Boeing 787 crash-tests a composite fuselage section was crashed to verify the accuracy of computer simulations. The report shows that the predictions were not top but still the tests showed principles to make the results predictable.
Remarkable is following structural concept: "The ring-frame configuration as commonly used in fixed-wing aircraft, is a difficult component for crashworthiness. NASA studies have shown, that the "point-load" applied by the ground leads to immediate fracture at this point, followed by severe bending, and further breakages of the frame higher up [7]. This may result in an early disintegration of the structure, and the bypassing of the dedicated energy-absorbing measures. The recommendations followed from these studies were, to separate the livable volume "on top" from the energy absorbing components "below". Hence, a sufficiently strong, closed ring-frame, meant to survive the impact, should be positioned on top of an expandable energy absorbing structure...."
In summary: Having a separated upper and lower structure in fuselage allows in theory a better design regarding crash-worthiness. And regarding the panel-advantage-discussion: By using panels instead of a barrel this concept can be implemented better.
The second link gives deep insight into European composite fuselage research: http://www.dlr.de/wf/en/Portaldata/2...t-fuer-einen-cfk-flugzeugrumpf.pdf
Document review:
Page 5 below: Sidenote about new metal technologies: they promise to have the same potential as today composites regarding cost and weight. But, future composites promise to be unbeatable. Therefore second generation composite fuselages will have a clear advantage over what is top today.
Page 6 below: Possible concepts to make composites conductive (hint: improved lightning protection)
Page 8 below: As above a separation of lower and upper fuselage functions.
Page 9 Top: Shell concept with integrated stringers. Design with co cured stringers and frames is achieved.
Page 11+12: Superior crash worthiness of Gondola concept
Page 12 below: Innovative and cheaper manufacturing, reduced number of bolts
Page 13 below: Summary of various European composite fuselage research programs
General: Work seems more concrete for an A320 sized fuselage, even a mixed demonstrator is mentioned on A320 basis. Building an A320 with a composite fuselage is therefore not impossible.
Now what do you think? Will the panels for the A350 already be superior to any barrel approach?
Forecasts
Based on all of this and to stir up the discussion I want to make some forecasts. These are only personal opinions and I will freely admit my wrong guess if the opposite turns to be the truth:
Forecast 1: CFRP Barrel fuselages will turn out to be inferior regarding crash worthiness
Forecast 2: CFRP Barrel fuselages will turn out to be more expensive in production
Forecast 3: CFRP Barrel fuselages will turn out to be heavier
Forecast 4: The B787 will be the last airliner using CFRP barrel fuselages
Forecast 5: The NG Narrowbodies from Boeing and Airbus will use CFRP panels
Forecast 6: The first 787 write off will be after an incident smaller than BA038 because of unrepairable fuselage damage (something like this:
, such a hole can not be patched)
Forecast 7: After the first comparing incident with an A350 the plane will not be written off
Forecast 8: CFRP Panel fuselages will allow replacement of single panels for repair of heavy damage (how?: Put the plane in a corset to fix the shape while part of the skin is missing, remove the fasteners for the panel, skin with stringers and frames is replaced as one piece).
AutoThrust From Switzerland, joined Jun 2006, 981 posts, RR: 4 Reply 1, posted (10 months 9 hours ago) and read 3788 times:
Very interesting post, Rheinwaldner. I'm no expert, but a other reason for shells are the need of smaller autoclave's unlike Boeing. You need gigantic autoclaves for Barrels.
Btw Airbus is doing research on a EU project called MAAXIMUS which include advancing simulation-based composite airframe design, improving manufacturing technology, and achieving a reduction of 10% in structural weight and 20% in development lead time, all by 2012.(could be applied on the A320RS)
HawkerCamm From United Kingdom, joined Jul 2007, 168 posts, RR: 0 Reply 2, posted (10 months 8 hours ago) and read 3697 times:
Excellent post and a very good technically analysis.
The mould lines being on the outside of the fuselage should also led to a very smooth external finish. The B787 barrels results in a fairly rough external finish
Overall some interesting remarks you make about the different methods of how to make the best Carbon type fuselage.
I don't want to comment on all conclusions you made but point 6 is certainly not true, even major damage as that (like a collision with a service truck or something similar) should be repairable when you use a barrel.
The patch (basically a custom made panel) could add some weight to the airframe but it should be minimal.
Tdscanuck From Canada, joined Jan 2006, 3759 posts, RR: 28 Reply 4, posted (10 months 6 hours ago) and read 3541 times:
Quoting Rheinwaldner (Thread starter): could there also be technological advantages that we just don't know?
I just mention in short the possible advantages that were listed so far:
* Easier logistics when transporting panels instead of barrels
* Longer panels than barrels possible
The logistics are definitely easier, but I don't believe the second point is true. The limit on length of panels or barrels today is the length of the autoclave...a 100' autoclave can only make a 100' panel or barrel. Diameter isn't as much of an issue as you might think since, for pressure reasons, autoclaves are almost always round anyway.
Quoting Rheinwaldner (Thread starter): With Panels the outside of the fuselage lay in the Mandrel. Therefore you will get the smother surface on the aerodynamically important side of the wall (or at least better control on the outcome during manufacturing).
There is no particular reason you can't do a barrel with a tooling surface on the outside, although I don't believe the 787 is doing so. I don't think this is an inherent panel vs. barrel issue.
Quoting Rheinwaldner (Thread starter): With barrels on the inside you can only vary the thickness to some degree and have limited possibilities to built 3D structures (for the B787 only stringers)
Provided you're willing to make the mandrel more complex, you can put structures and thickness changes to almost arbitrary degree for panels or barrels.
Quoting Rheinwaldner (Thread starter): #
# I made a graphic that shows a panel with fully integrated stringers and frames:
There's a slight issue here with your lap joint...tremendous off-center loading (and a heck of a stress riser in the sharp corner). Although I certainly agree you could integrate the lap joint into the panel, it strongly doubt the cross section would look like this.
Quoting Rheinwaldner (Thread starter): The panel approach leads to higher integrated parts allowing cheaper (more automated) production and final assembly. There are fewer fasteners and joints (if you count the joint between each frame and the barrel too).
Both panel and barrel construction can support higher integration than either Airbus or Boeing are currently considering. I'm don't see why your point above applies to panels and not barrels.
Quoting Rheinwaldner (Thread starter): Having a separated upper and lower structure in fuselage allows in theory a better design regarding crash-worthiness. And regarding the panel-advantage-discussion: By using panels instead of a barrel this concept can be implemented better.
Again, not sure why. One of the huge advantages of composites, in both panel and barrel form, is local tailoring of structural properties. You can build a barrel with very different upper and lower crash response just as you can build it with panels.
Quoting Rheinwaldner (Thread starter): Forecast 1: CFRP Barrel fuselages will turn out to be inferior regarding crash worthiness
As a general statement, I believe this is false because of the point raised above.
Quoting Rheinwaldner (Thread starter): Forecast 2: CFRP Barrel fuselages will turn out to be more expensive in production
Absolutely true on a part basis. For an overall airplane cost basis, I doubt it because of the savings in final assembly joints.
Highly doubtful given that, no matter how you slice it, panels will have more major joints.
Quoting Rheinwaldner (Thread starter): Forecast 6: The first 787 write off will be after an incident smaller than BA038 because of unrepairable fuselage damage
Doubtful. Repairability is negligably different between the two technologies because for anything up to a full panel replacement, you're going to be doing the same patch on either one. If you go to full panel replacement (a *major* repair) the equivalent patch on a barrel is of the same complexity.
Quoting Rheinwaldner (Thread starter): Forecast 8: CFRP Panel fuselages will allow replacement of single panels for repair of heavy damage (how?: Put the plane in a corset to fix the shape while part of the skin is missing, remove the fasteners for the panel, skin with stringers and frames is replaced as one piece).
This is certainly true, and is equivalent to a reskinning job today. However, this is a huge repair and isn't undertaken lightly. A tiny fraction of a percent of all fuselage repairs over the life of an entire fleet are done by skin replacement, so the value amortized over the fleet is relatively small.
RedFlyer From United States, joined Feb 2005, 3175 posts, RR: 20 Reply 5, posted (10 months 4 hours ago) and read 3486 times:
Very interesting thread.
Wouldn't the Panel Max approach actually increase manufacturing costs? It sure looks "prettier" and cleaner from a fastener standpoint, but I imagine co-curing the stringers, frames, doorframes, and window-frames is going to add a considerable level of complexity to the manufacturing process. The forms (press) are going to have to be very precise within very tight tolerances and all the while accounting for a multitude of complex shapes and angles. More to the point, would it even be possible to cure a panel containing all of those different parts at one time or would the panel have to undergo multiple curings? Wouldn't one also require a unique form or press for each panel? Would that also entail an autoclave for each form or press?
I know, a lot of questions. I'm just an ignorant layman, but that is what popped into my mind when I read throught the OP's post.
"I am looking for the owner of that horse - he's tall, he's blonde, he smokes a cigar, and he's a pig."
Tdscanuck From Canada, joined Jan 2006, 3759 posts, RR: 28 Reply 6, posted (10 months 2 hours ago) and read 3462 times:
Quoting RedFlyer (Reply 5): Wouldn't the Panel Max approach actually increase manufacturing costs?
As compared to a regular panel, probably, although it would be offset somewhat by reduced assembly costs.
Quoting RedFlyer (Reply 5): More to the point, would it even be possible to cure a panel containing all of those different parts at one time or would the panel have to undergo multiple curings?
Unlikely, due to handling problems if nothing else. It would be easier to partially cure the parts individually to give them some structure and allow handling without them distorting, then do the final co-cure to bond it all together.
Quoting RedFlyer (Reply 5): Wouldn't one also require a unique form or press for each panel?
Unique tooling or reconfigurable tooling. Probably no press involved (that's what the autoclave is for).
Quoting RedFlyer (Reply 5): Would that also entail an autoclave for each form or press?
Probably not...the autoclave is just a big pressurized oven. Provided it's big enough for your parts and tooling to fit in, you can use it over and over on different parts (or cure multiple parts together, if the cure cycle is the same).
RedFlyer From United States, joined Feb 2005, 3175 posts, RR: 20 Reply 7, posted (9 months 4 weeks 1 day 19 hours ago) and read 3408 times:
Quoting Tdscanuck (Reply 6): Quoting RedFlyer (Reply 5):
More to the point, would it even be possible to cure a panel containing all of those different parts at one time or would the panel have to undergo multiple curings?
Unlikely, due to handling problems if nothing else. It would be easier to partially cure the parts individually to give them some structure and allow handling without them distorting, then do the final co-cure to bond it all together.
That was my whole point: "partially curing the parts individually...then do the final co-cure" would seem to be multiple curings. All of which adds time and complexity to the process. It certainly doesn't make it more efficient. The individual parts will still have to be "constructed" prior to the final sub-assembly. Rather than construct the support structures on the back-end of the process as Boeing does on its Barrrel method, Airbus would construct the support structures on the front-end of their Panel method.
Quoting Tdscanuck (Reply 6): Quoting RedFlyer (Reply 5):
Wouldn't one also require a unique form or press for each panel?
Unique tooling or reconfigurable tooling. Probably no press involved (that's what the autoclave is for).
Just curious, but how would the parts be held in place during the autoclave process? I was under the impression an autoclave applies pressure and heat to "cure" the compound by removing all air from the material. How would the separate parts be applied (all at the same time, I presume) under pressure while ensuring they remain static? (That was why I used the term "press".)
Quoting Tdscanuck (Reply 6): Quoting RedFlyer (Reply 5):
Would that also entail an autoclave for each form or press?
Probably not...the autoclave is just a big pressurized oven. Provided it's big enough for your parts and tooling to fit in, you can use it over and over on different parts (or cure multiple parts together, if the cure cycle is the same).
I could see using a single autoclave for different panel parts. But that would kind of slow things down a little, wouldn't it? I believe we're talking roughly a dozen panels that would comprise the fuselage. That's 12 individual "cooks" that have to take place. Seems it would remove a lot of the efficiency if that has to be done (and we won't even go into defective panels that have to be re-done). Unless, of course, Airbus were to outsource the panels to different vendors who could cure panels at the same time at different locations.
In any event, the OP's original post is eye-opening. I can see some of the advantages to the Panel approach (the Panel Max option). I don't know if it's superior than the Barrel method, but definitely better than I thought it was before.
"I am looking for the owner of that horse - he's tall, he's blonde, he smokes a cigar, and he's a pig."
WingedMigrator From United States, joined Oct 2005, 1503 posts, RR: 27 Reply 8, posted (9 months 4 weeks 1 day 18 hours ago) and read 3401 times:
Thanks Rheinwaldner for the interesting post. Your picture 3 "panel max" is indeed what I had been inquiring about, without success, and you explained it and illustrated it much more clearly.
I don't follow your conclusions on the superiority of the panel approach... in my view there are too many variables and unknowns that prevent the categorical statement that one or the other is better.
But you are correct that the barrel approach, being just one of several possible approaches, is sometimes oversold. Or, more to the point, the panel approach is often dismissed as inefficient, adding what sounds like many tons -- when in fact these bolted joints are a tiny fraction of the mass of the finished airplane. All this is very far from being clear at this point.
Quoting RedFlyer (Reply 5): I imagine co-curing the stringers, frames, doorframes, and window-frames is going to add a considerable level of complexity to the manufacturing process.
Yes... but Airbus seems to have some experience in this regard. For example, the A380 rear pressure bulkhead (or rather, Bulkhead with a big 'B') is a monolithic CFRP structure that undergoes at least 2 cure cycles, one for the skin and the second for co-curing ribs. Perhaps some of this complexity could be offset by the simpler assembly.
Quoting RedFlyer (Reply 7): The individual parts will still have to be "constructed" prior to the final sub-assembly.
By machines, perhaps? I'm not sure why this would be a disadvantage.
Quoting RedFlyer (Reply 7): I believe we're talking roughly a dozen panels that would comprise the fuselage. That's 12 individual "cooks" that have to take place.
You are possibly assuming that only one panel is allowed per autoclave cycle. I see nothing in the panel approach that prevents a rack of several panels from being cured simultaneously.
Quoting RedFlyer (Reply 7): and we won't even go into defective panels that have to be re-done
That's potentially a problem with barrels as well, since losing a barrel is like losing four panels at once. If the yield were significantly less than 100%, panels would be advantageous, but everything I've heard about the Boeing operation suggests that yield is very close to 100%.
RedFlyer From United States, joined Feb 2005, 3175 posts, RR: 20 Reply 9, posted (9 months 4 weeks 1 day 15 hours ago) and read 3380 times:
Thanks for keeping this thread going, WingedMigrator. I find it one of the more interesting so I'm looking for more detailed explanations. On to your specific points...
Quoting WingedMigrator (Reply 8): Quoting RedFlyer (Reply 5):
I imagine co-curing the stringers, frames, doorframes, and window-frames is going to add a considerable level of complexity to the manufacturing process.
Yes... but Airbus seems to have some experience in this regard.
No doubt Airbus has experience with this and they could do it very efficiently. What I'm curious to know is how this specifically might be a better (read more efficient) process than Boeing's barrel approach. As I said, Boeing is constructing the stringers, frames, door frames, and window frames on the back-end of the construction process. So Airbus does it on the front end - how does that make it more efficient? Parity, I can understand. I'm not sure I grasp the "more" efficient aspect of it.
Quoting WingedMigrator (Reply 8): Quoting RedFlyer (Reply 7):
The individual parts will still have to be "constructed" prior to the final sub-assembly.
By machines, perhaps? I'm not sure why this would be a disadvantage.
I would think a lot of the sub-components (stringers, frames, doorframes, and window frames) are also machined on the Boeing side. Don't forget: most of the parts and sub-assemblies by both manufacturers are machine (robot) made these days.
Quoting WingedMigrator (Reply 8): Quoting RedFlyer (Reply 7):
I believe we're talking roughly a dozen panels that would comprise the fuselage. That's 12 individual "cooks" that have to take place.
You are possibly assuming that only one panel is allowed per autoclave cycle. I see nothing in the panel approach that prevents a rack of several panels from being cured simultaneously.
Ah, definitely a big step forward! A large-enough autoclave could have several "shelves" or racks that accommodate multiple panels and that would definitely make it a more efficient process.
Quoting WingedMigrator (Reply 8): Quoting RedFlyer (Reply 7):
and we won't even go into defective panels that have to be re-done
That's potentially a problem with barrels as well, since losing a barrel is like losing four panels at once. If the yield were significantly less than 100%, panels would be advantageous, but everything I've heard about the Boeing operation suggests that yield is very close to 100%.
Ok, I see your point; however, in manufacturing "defects" are defined by defects per "X" number of parts. If we assume all of the sub-components (stringers, frames, etc.) are a wash between the two methods since they will both require manufacture of those components in similar quantities separately for inclusion in the final subassembly products, then the barrel method actually has LESS parts than the panel method. That's because currently Boeing constructs 7 barrels per fuselage (I believe their goal is to eventually reduce this number to 4 or even 3 barrels). The panel method would require roughly a dozen panels (parts) for a fuselage. If we are to assume 1 defect per 100 parts then that means that every 1 of every 8 airframes will encounter a defective part requiring a re-do. On the barrel method, however, 1 defect per 100 parts translates into 1 defective part for every 14 airframes. (This is sounding an awful lot like that age-old A.net argument about twins vs. quads!) Of course, I'm assuming 1 in 100 defects - it could in fact be a lot better ratio which could render it a relatively moot point. And, of course, people could point out that a mistake on a Boeing barrel part, while less likely, is more costly because it's a bigger part (more materials, more time to construct, etc.).
(BTW, neither of the links you provided worked on my end.)
[Edited 2008-01-26 00:00:56]
"I am looking for the owner of that horse - he's tall, he's blonde, he smokes a cigar, and he's a pig."
WingedMigrator From United States, joined Oct 2005, 1503 posts, RR: 27 Reply 10, posted (9 months 4 weeks 1 day 5 hours ago) and read 3318 times:
Quoting RedFlyer (Reply 9): What I'm curious to know is how this specifically might be a better (read more efficient) process than Boeing's barrel approach.
I haven't claimed that one was superior to the other. In my view the pros and cons of each method are not sufficiently well-defined to conclude positively one way or the other. Others do conclude... Rheinwaldner is in the 'panel' column, and lots of 787 fans are firmly in the 'barrel' column, but I don't think either side has made a convincing case that a clear advantage exists.
Quoting RedFlyer (Reply 9): As I said, Boeing is constructing the stringers, frames, door frames, and window frames on the back-end of the construction process.
Boeing's stringers are integral to the barrels. Everything else is, as you say, manually installed inside the barrels with fasteners. Co-curing all that at the front end of the process may further cut down on the amount of touch labor. The panel approach may eliminate as many fastened joints as it adds.
Quoting RedFlyer (Reply 9): in manufacturing "defects" are defined by defects per "X" number of parts.
One thing to keep in mind is that "X" probably depends on the size of the part. This is true of microchips and computer displays: the larger the part, the lower the yield, because the defects occur at a given rate on a scale much smaller than the whole part. I don't know if this is true of large CFRP parts-- as I said, nothing indicates that Boeing's yield is suffering from the large size of their parts.
Quoting RedFlyer (Reply 9): neither of the links you provided worked on my end
That's really too bad. This link from 2003 has a fascinating, detailed and illustrated description of the manufacturing steps for the A380 bulkhead. They mention that eventually, reinforcing features could be co-cured in a single autoclave cycle, but they are sticking with two cycles for now because they want to stay within their experience and lower risk.
It's not too difficult to imagine a similar manufacturing process for the A350 fuselage panels, in the "Panel Max" configuration described by Rheinwaldner above.
474218 From United States, joined Oct 2005, 2818 posts, RR: 3 Reply 11, posted (9 months 4 weeks 1 day 3 hours ago) and read 3303 times:
Quoting Rheinwaldner (Thread starter): I made a graphic that shows a panel with fully integrated stringers and frames:
It would come as one piece out of the oven. The number of fasteners is greatly reduced. The splice plate on one side could be made from the panel itself thus eliminating further rows of fasteners.
As you have it shown, you have rivets in single shear and that is not acceptable for primary structure. There would have to have an internal and/or an external butt strap to add structural integrity. A rivet through three layers of material (double shear) has over twice the shear strength as a rivet through two layers (single shear).
Sponsor Message: 
From Switzerland, joined Jan 2008, 363 posts, 
From United Kingdom, joined Jul 2007, 168 posts, 
From Belgium, joined Aug 2005, 888 posts, 
From Canada, joined Jan 2006, 3759 posts, 
From United States, joined Feb 2005, 3175 posts,