In general, flying higher is better (low wing loading)
AFAICS this is true on account of 2 dynamics:
- 1. Parasitic drag is usually predominant and is linear with air density (induced drag, usually smaller than parasitic, is inversely linear with density).
- 2. Engines run more efficiently in colder air.
Neither condition, however, really favors flying above FL37 for the A380.
Re (1), Di's predominance for A380 means its L/D actually decreases as it ascends.
Re (2), above FL37 or so the stratosphere is isothermic and the engine efficiency benefit runs out.
A380 reaches FL37 pretty quickly, after which ascent only means lower true air speed (due to compressibility effects) and slightly worse L/D.
I wonder if a longer 6+8 or 6+9 would solve the empennage weight problem while also being tall enough to counteract bending moments.
It probably would be tall enough: 8/9+6 is only slightly less tall than a good 10+8 because deck/hold height sets the vertical dimension. The slightly longer fuse helps with empennage size too of course.
Here's a confession: I haven't been completely honest about my criticism of 10+8 in preference to 8/9+6 for new design. When I've done a little modeling, a good 10+8 can pencil out better than 8/9+6 even for 500 seats. The key is that, as mentioned, deck height controls the vertical dimension so additional width comes at little cost in Swet. Total fuse Swet is, as PolarRoute
suggests, meaningfully more efficient with 10+8. And a less oblong cross-section gives some small benefit in Cdp.
My lack of candor owes to laziness rather than trying to pull one over on y'all: explaining at length would have required more work and more anti-A380 invective. Basically the A380's 10-8 is an extremely inefficient 2-deck cross-section (despite being by the most efficient cross section ever actually built). Excessive MD and belly height is a main culprit (motivated by cargo per Father A380). Equally important, max cabin width is up around MD shoulder level. Plus the extreme sidewall thickness that a modern 10+8 design would ameliorate with 777X-style sidewall shaving at key points and/or thinness enabled by CFRP. A double-bubble is far more efficient for cabin space but less so for bending resistance. I'd guess that's the main reason for not going double-bubble on the A380: if you need to stretch it to 80m the bending stress becomes more dominant.
But it's not just the cross section: A380's fuse Swet is further harmed by the 43ft empty tailcone. For comparison, the MD-12 design had a ~21ft empty tailcone and similar large cross section. So fuse height alone can't explain that tailcone length. Rather it seems the massive tailcone is necessary to fit the massive tailplane, whose massive elevation machinery sits ahead of the massive tailplane. The more I learn about this program, the less surprised I'd be to find out the tailcone is massive because, "meh, it'll sell anyway because it's big." I.d. Airbus just wasn't trying very hard.
In sum I'd say a new 2-decker is probably optimal at 10-8 for 500 seats, maybe not for 450 seats but probably still is. As 450 seats seems around the "kink" where the capacity-efficiency curve transitions from single to double decker, we can see very significant variations in efficiency at this point between double-deckers. So the MCC of a 10-8 500-seater may be attractive, compared to a 8/9-6 450-seater. I'm far less confident about that judgment though, would require more resolution than I have from the fundamentals.
For the A380, however, anything that forced Airbus to abandon the 650-seat target would have been an incredible boon to the program, even if it turned out that 8/9-6 was slightly less efficient than a notional 10-8. That's because re A380 we're not analyzing a notional 10-8, we're analyzing the actual A380 cross section that wasted space/weight/drag in the vertical axis and placed max width at a useless elevation. A good 8-6 would have been far better than the A380 but not necessarily because 8-6 is inherently better than 10-8.
Another difference between analyzing a contemporary optimized 2-decker versus analyzing alternate A380 designs is the differing predominance of the empennage factor. For equal capacity and range, an optimal 2000 design will have a larger empennage than an optimal 2025 design (higher operating weights and bigger engines means bigger control surfaces). The older design would place more emphasis on fuse lever arm to minimize empennage size than would the newer. So 8-6 could have been optimal in 2000 without implying it's optimal now.
Why do you think it needs to have >1 span/length?
The reasoning incorporates the above analysis of 10-8 vs. 8/9+6. If your optimization loop spits out the capacity of an 80m 8/9+6 you've already got something larger than A388: cabin length is going to be ~230ft vs. ~165 (IIRC) for A388 because you reclaim empty tailcone space in the optimal design. So even if 8+6's combined cabin width is only 70% of A380's two decks, total deck space is at least as big (aside from longer, less of the 8+6 cabin is tapered). At that size (~550 seats) I'm fairly certain that 10+8 is more efficient overall, even with a slightly bigger empennage than 8+6 (again for contemporary design, not necessarily for 2000). If you want smaller capacity then it's not 80m length. Maybe the wing shrinks in line with fuselage length but 80m seems optimal for contemporary design even if the 8-6 fuselage is only 240ft long. If you pencil around with 2025 projections of an optimized 2-decker, it becomes clear that payload weight is more significant than usual. So you need a big wing to carry lots of people and any cargo. There's diminishing returns to structural efficiency unless you can discriminate against fat passengers (maybe a small-wing subvariant for East Asia?).
Do you mean compared to the average B777-300ER? 40% sounds a lot.
Can you expand?
It's a lot but start from saving 40% of fuselage drag per pax and work your way around the fundamentals. Add in Ultrafan and it's conservative, IMO.
Google matt6461 and "A380NEO," "A380X," and "A380NWO" for my old threads in TechOps. The older a thread is, the less I knew when I wrote it. The A380-900NEO probably has my least ignorant analysis but in each thread the fundamentals discussions are fairly accurate.