6 ms·

Why airplane windows have round corners (2016)

133 points by smacktoward 7y ago

perl4ever 7y ago
Interestingly, the fuselage fragment picture looks like the windows were rounded. Also, while I have read the story about the Comet many times before, I vaguely recall one telling of it that noted the problem wasn't simply due to square windows, but some other contributing factor, an adjustment to the design that caused extra stress on top of the corners. It kind of seems like common sense that engineers wouldn't have been as oblivious as the standard telling implies. Most disasters don't happen from anyone being dumb per se, but from the interaction of different people or groups who lack the full context of what the other is doing, and I believe that was the case here.
- stcredzero 7y ago
  I vaguely recall one telling of it that noted the problem wasn't simply due to square windows, but some other contributing factor, an adjustment to the design that caused extra stress on top of the corners. An opening on the top for a radio receiver was improperly riveted, instead of being glued. The rivet holes caused stress concentration, which resulted in metal fatigue and eventual failure and explosive decompression. When this root cause was discovered, it was also realized that the window corners also caused stress concentration.
- WalterBright 7y ago
  Not rounded enough. There are formulas for the stress on an inside corner, and the larger the radius, the lower the stress rise.
pvaldes 7y ago
Ok, is interestiong, so lets eval the explossive decompression issue. What if we could lead all this air to an external compensating device in a more ordered and less violent way?. I'm thinking on some kind of "external airbag", elastic membrane, rubber balloon, etc, so we would reduce the time of event duration and soften or control the disordered exit of air (and maybe to save also some valuable oxygen from being totally lost). If unfortunately the accident happens, any corpse, object or piece would be trapped in the balloon instead to be sucked out and dissapear (this is better for the family than not having nothing to bury at least). I suppose that to implement such device, like a rubber ring window in a handful of windows and doors, wouldn't be technically defiant and could save lifes (or alleviate the harm done) until emergency landing. Would be also a inmediately visible flag for airport rescue team to go directly to the area with more possible victims.
- dimator 7y ago
  You're solving a problem that doesn't exist anymore.
  - pvaldes 7y ago
    It seems that decompressive explosion and people killed after being sucked by a broken window is still an unprobable but lethal risk, unfortunately https://en.wikipedia.org/wiki/Southwest_Airlines_Flight_1380 https://en.wikipedia.org/wiki/Southwest_Airlines_Flight_1380
    - rtkwe 7y ago
      It's not really a risk that an air bladder or anything like what OP was suggesting will fix though. It was powerful enough to break the window it would probably be powerful enough to break the bladder and the window. It's generally better to just fix the direct problem instead of a weird patch to catch the failure.
      - pvaldes 7y ago
        A hammer will break first a glass ball than a gel ball or a rubber tire. We know how to build solid airplanes made 100% with rigid materials except in a single point (We know that rubber tires are the right way to cope with repeated landing impacts). Maybe we should explore a different way and include more deformable materials in strategic points to absorb the damage made by a possible impact. Maybe the concept of an "boxer's eyelid airbag" for windows in airplanes, would deserve some exploratory research.
- dredmorbius 7y ago
  Venting from multiple, engineered, locations rather than a single explanding rupture would spare the airframe from catastropic disintegration. It's still hell on the cargo and pilots. You'd need a venting system which itself would detect sudden pressure change and blow first -- effectively a pressure circuit breaker. That itself would all but certainly introduce multiple new catastrophic failure modes. Your suggestion generalises to flying an airplane's pressure hull within another pressure hull. This scales poorly in a regime where weight and payload space are highly constrained (though it has similarities to nonpressurised oil tanker double-hull design). Yo dawg, I hear you like airplanes, here's an airplane you can put your airplane in so you can fly your airplane in an airplane.... Better to design-proof against metal fatigue-induced ruptures in the first place.
tyingq 7y ago
Wikipedia says the crashes weren't caused by the windows: "The accident report's use of the word "window" when referring to the Automatic Direction Finding (ADF) aerial cutout panel[121] has led to a common belief that the Comet 1's accidents were the result of its having square passenger windows." https://en.wikipedia.org/wiki/De_Havilland_Comet https://en.wikipedia.org/wiki/De_Havilland_Comet
- starpilot 7y ago
  This seems to contradict other parts of the Wikipedia article saying the windows were the problem: > Design and construction flaws, including improper riveting and dangerous concentrations of stress around some of the square windows, were ultimately identified. As a result, the Comet was extensively redesigned, with oval windows, structural reinforcements and other changes. There's an ongoing debate on this on the article's talk page.
  - tyingq 7y ago
    I read that as solving a potential future issue, versus solving the issue that caused the crashes. Here's a diagram of what the ADF "windows" are: https://images.app.goo.gl/3AUW6JsPV6tj39fq9 https://images.app.goo.gl/3AUW6JsPV6tj39fq9
  - WalterBright 7y ago
    Boeing windows are not only rounded, there's a forged doubler riveted around the rim. It's kinda heavy, which is another reason why cargo versions don't have windows.
ken 7y ago
> "A jet airliner, in other words, was an airliner that could revolutionize aviation — and that meant the first company to bring one to market stood to make an absolutely fantastic amount of money." Good on him for leading the innovation, but it seems obvious to me today that "first" rarely means "most profitable" or "longest surviving".
_bxg1 7y ago
> Too late for the British aerospace industry, which saw competitors like Boeing and Lockheed, slower to market with their own jetliners but perceived to be more safe, seize and hold the crown of king of the Jet Age for America. Irony.
jessriedel 7y ago
The only thing worse than not having a tl;dr is having a clickbait tl;dr: "The answer is simple: to keep that window from killing you". (Translation: "To find out more, read the article!") The actual claimed explanation: > when an aircraft’s interior is pressurized and de-pressurized repeatedly, over and over again for many months, the strength of that aircraft’s metal body slowly weakens — a phenomenon that became known as metal fatigue. And when the holes you cut into that body to hold windows have sharp corners like squares do, thanks to a process called stress concentration the weakness builds up much faster in those sharp corners than it does elsewhere You can read the rest of the article if you want to read about the history of the de Havilland Comet, the world's first commercial jet airliner.
JulianMorrison 7y ago
So, why aren't they circular like portholes?
- em-bee 7y ago
  well, if you look at the first photo in the article, they actually were completely circular. i suppose eventually people decided to compromise and found out how much roundness is enough.
- taneq 7y ago
  You can probably get 80% of the stress reduction from 20% circularity.
  - Gibbon1 7y ago
    Definitely that. Also for a cylinder under pressure the hoop stress is twice the longitudinal stress. Which says to me that vertically oriented rectangular windows are optimal. Probably even more so since the windows are aligned.
gmueckl 7y ago
Stress concentration, especially at sharp corners, is a well known phenomenon. During the second world war, the US manufactured transport ships in an assembly line fashion. The were launching these ships at rates of more than one per week. A lot of them had the same design flaw: the cargo hatch corners were not properly rounded off. This made them starting points for cracks in the hull. Amd once a crack is started, it continues to extend relentlessly under stress. As the story goes, a few of these ships were lost because the cracks grew so long that the hulls broke and sank.
nutcracker46 7y ago
It is unfortunate that people lost in the Comet crashes could not have enjoyed rides on better engineered aircraft, but that is how advancement really happens. Look at the DC-10 and B-737 Max. As to secondary walls in the pressure vessel, bladders, etc., take a look at actual designs for both windows and cabin pressure outflow valves are designed. They work well and can withstand quite a bit of adverse conditions. Those outflow valves are complex - usually consisting of motorized doors plus one or two spring loaded check valves for overpressure events. Usually there is another outflow valve for purging smoke or creating extra cooling flow for electronics under the flight deck. Going back to windows and other cutouts, there just wasn't enough sharing of expertise among ship and aircraft builders. There wasn't deep enough thought about effects of pressurization cycles. Now we know.
jjakque 7y ago
Real Engineering made a video on this exact topic, explaining the essence in less than 2 minutes: https://www.youtube.com/watch?v=7rXGRPMD-GQ https://www.youtube.com/watch?v=7rXGRPMD-GQ He went through bit more details said topic in another video "Why Are The Dreamliner's Windows So Big": https://www.youtube.com/watch?v=7-I20Ru9BwM https://www.youtube.com/watch?v=7-I20Ru9BwM
NamTaf 7y ago
Stress concentration factors like this are the bane of all structural engineering. Anything with too sharp an internal corner will concentrate stress and lead to material fatigue failure if there's any stress there. Given I've spent all morning staring at it in some finite element analysis work, I'm going to ramble about it to give myself a break. Sometimes dealing with it can be a pain, for example if you need two rectangular things to sit together along two faces, since you have to radius both corners with a larger radius on the smaller object. One way to combat this is to drill a hole in the corner, removing additional material but creating a little 270ish-degree-C-shaped radius in place of the sharp corner. Take what would've originally been a sharp corner, but then drill it out (example: [1]). What you lose on less material, you more than gain back with eliminating the stress concentration. This has the advantage that then a rectangular object can fit snug into the corner, because the smaller's sharp corner just sits inside the larger's C-shape cutout. The best visualisation tool of this phenomenon that I learned is force lines [2]. Where they bunch up closely together is where high stress occurs. It follows naturally that a sudden sharp corner will concentrate all those lines close to it, whereas a gentler radius encourages a more gradual transition. Think of the lines as similar to Isobars (probably the most well-known contour line) [3], in that the lines represent points of constant pressure/force, and so will sort-of repel adjacent lines (since there needs to be a smooth gradient between those lines). The more gradually they're forced to change, the more gradual that gradient between lines will be (represented by them being farther apart). Anyhow, all of this is also why you see countless concrete footpaths, driveways, etc. with cracks growing out from sharp corners in the concrete. Look at any internal corners where the concrete runs in an L shape. You'll not uncommonly find a crack growing from it. However, no one really cares (except for aesthetic properties) because it doesn't cause a functional problem, and forming concrete into straight sides is orders of magnitude easier than forming smooth curves. Bonus points if you notice that it's generally not parallel to one of the edges that form the corner, and can figure out why :) (I'd be guessing if I gave my answer, but I'm somewhat confident it'd be right: thermal expansion/contraction affects both sides simultaneously, meaning that the shear stress runs generally at 45 degrees to the corner rather than perpendicular to one, which would be the case if only one side expanded/contracted.) [1]: https://img3.gmdu.net/35780.0.jpg https://img3.gmdu.net/35780.0.jpg [2]: https://en.wikipedia.org/wiki/Force_lines https://en.wikipedia.org/wiki/Force_lines [3]: https://en.wikipedia.org/wiki/Contour_line#Barometric_pressure https://en.wikipedia.org/wiki/Contour_line#Barometric_pressu...
ganesh7 7y ago
You could also say the great innovators did not understand the technology they looted from the Germans. Or the similarities in design, e.g. integration of turbines in the wings, of the Comet and Nazi jet planes are merely coincidental I assume.
- dTal 7y ago
  The placement of the engines has nothing to do with metal fatigue from thousands of repeated 40,000 feet cruises with a pressurized cabin anyway; no German jet fighter was pressurized, so it's doubtful that they understood the technology either. Also, the only Nazi aircraft of any type with root-integrated jet engines is the obscure Horton 229 prototype - and apart from that single cosmetic similarity, it has absolutely nothing in common with the de Havilland Comet.
  - ganesh7 7y ago
    Uhuh: "The Junkers Ju 49 was a German aircraft designed to investigate high-altitude flight and the techniques of cabin pressurization. It was the world's second working pressurized aircraft, following the Engineering Division USD-9A which first flew in the United States in 1921.[1] By 1935, it was flying regularly to around 12,500 m (41,000 ft)."
    - dTal 7y ago
      What on Earth is your argument? Everyone was experimenting with cabin pressurization - that wasn't even the first one. Are you seriously suggesting that that specific aircraft - a pre-war, single-engine piston craft - in any way influenced the design of the Comet? https://en.wikipedia.org/wiki/Cabin_pressurization#History https://en.wikipedia.org/wiki/Cabin_pressurization#History

133 points by smacktoward 7y ago

Interestingly, the fuselage fragment picture looks like the windows were rounded. Also, while I have read the story about the Comet many times before, I vaguely recall one telling of it that noted the problem wasn't simply due to square windows, but some other contributing factor, an adjustment to the design that caused extra stress on top of the corners. It kind of seems like common sense that engineers wouldn't have been as oblivious as the standard telling implies. Most disasters don't happen from anyone being dumb per se, but from the interaction of different people or groups who lack the full context of what the other is doing, and I believe that was the case here.

stcredzero 7y ago
I vaguely recall one telling of it that noted the problem wasn't simply due to square windows, but some other contributing factor, an adjustment to the design that caused extra stress on top of the corners. An opening on the top for a radio receiver was improperly riveted, instead of being glued. The rivet holes caused stress concentration, which resulted in metal fatigue and eventual failure and explosive decompression. When this root cause was discovered, it was also realized that the window corners also caused stress concentration.
WalterBright 7y ago
Not rounded enough. There are formulas for the stress on an inside corner, and the larger the radius, the lower the stress rise.

pvaldes 7y ago

Ok, is interestiong, so lets eval the explossive decompression issue. What if we could lead all this air to an external compensating device in a more ordered and less violent way?. I'm thinking on some kind of "external airbag", elastic membrane, rubber balloon, etc, so we would reduce the time of event duration and soften or control the disordered exit of air (and maybe to save also some valuable oxygen from being totally lost). If unfortunately the accident happens, any corpse, object or piece would be trapped in the balloon instead to be sucked out and dissapear (this is better for the family than not having nothing to bury at least). I suppose that to implement such device, like a rubber ring window in a handful of windows and doors, wouldn't be technically defiant and could save lifes (or alleviate the harm done) until emergency landing. Would be also a inmediately visible flag for airport rescue team to go directly to the area with more possible victims.

dimator 7y ago
You're solving a problem that doesn't exist anymore.
- pvaldes 7y ago
  It seems that decompressive explosion and people killed after being sucked by a broken window is still an unprobable but lethal risk, unfortunately https://en.wikipedia.org/wiki/Southwest_Airlines_Flight_1380 https://en.wikipedia.org/wiki/Southwest_Airlines_Flight_1380
  - rtkwe 7y ago
    It's not really a risk that an air bladder or anything like what OP was suggesting will fix though. It was powerful enough to break the window it would probably be powerful enough to break the bladder and the window. It's generally better to just fix the direct problem instead of a weird patch to catch the failure.
    - pvaldes 7y ago
      A hammer will break first a glass ball than a gel ball or a rubber tire. We know how to build solid airplanes made 100% with rigid materials except in a single point (We know that rubber tires are the right way to cope with repeated landing impacts). Maybe we should explore a different way and include more deformable materials in strategic points to absorb the damage made by a possible impact. Maybe the concept of an "boxer's eyelid airbag" for windows in airplanes, would deserve some exploratory research.
dredmorbius 7y ago
Venting from multiple, engineered, locations rather than a single explanding rupture would spare the airframe from catastropic disintegration. It's still hell on the cargo and pilots. You'd need a venting system which itself would detect sudden pressure change and blow first -- effectively a pressure circuit breaker. That itself would all but certainly introduce multiple new catastrophic failure modes. Your suggestion generalises to flying an airplane's pressure hull within another pressure hull. This scales poorly in a regime where weight and payload space are highly constrained (though it has similarities to nonpressurised oil tanker double-hull design). Yo dawg, I hear you like airplanes, here's an airplane you can put your airplane in so you can fly your airplane in an airplane.... Better to design-proof against metal fatigue-induced ruptures in the first place.

tyingq 7y ago

Wikipedia says the crashes weren't caused by the windows: "The accident report's use of the word "window" when referring to the Automatic Direction Finding (ADF) aerial cutout panel[121] has led to a common belief that the Comet 1's accidents were the result of its having square passenger windows." https://en.wikipedia.org/wiki/De_Havilland_Comet https://en.wikipedia.org/wiki/De_Havilland_Comet

starpilot 7y ago
This seems to contradict other parts of the Wikipedia article saying the windows were the problem: > Design and construction flaws, including improper riveting and dangerous concentrations of stress around some of the square windows, were ultimately identified. As a result, the Comet was extensively redesigned, with oval windows, structural reinforcements and other changes. There's an ongoing debate on this on the article's talk page.
- tyingq 7y ago
  I read that as solving a potential future issue, versus solving the issue that caused the crashes. Here's a diagram of what the ADF "windows" are: https://images.app.goo.gl/3AUW6JsPV6tj39fq9 https://images.app.goo.gl/3AUW6JsPV6tj39fq9
- WalterBright 7y ago
  Boeing windows are not only rounded, there's a forged doubler riveted around the rim. It's kinda heavy, which is another reason why cargo versions don't have windows.

ken 7y ago

> "A jet airliner, in other words, was an airliner that could revolutionize aviation — and that meant the first company to bring one to market stood to make an absolutely fantastic amount of money." Good on him for leading the innovation, but it seems obvious to me today that "first" rarely means "most profitable" or "longest surviving".

_bxg1 7y ago

> Too late for the British aerospace industry, which saw competitors like Boeing and Lockheed, slower to market with their own jetliners but perceived to be more safe, seize and hold the crown of king of the Jet Age for America. Irony.

jessriedel 7y ago

The only thing worse than not having a tl;dr is having a clickbait tl;dr: "The answer is simple: to keep that window from killing you". (Translation: "To find out more, read the article!") The actual claimed explanation: > when an aircraft’s interior is pressurized and de-pressurized repeatedly, over and over again for many months, the strength of that aircraft’s metal body slowly weakens — a phenomenon that became known as metal fatigue. And when the holes you cut into that body to hold windows have sharp corners like squares do, thanks to a process called stress concentration the weakness builds up much faster in those sharp corners than it does elsewhere You can read the rest of the article if you want to read about the history of the de Havilland Comet, the world's first commercial jet airliner.

JulianMorrison 7y ago

So, why aren't they circular like portholes?

em-bee 7y ago
well, if you look at the first photo in the article, they actually were completely circular. i suppose eventually people decided to compromise and found out how much roundness is enough.
taneq 7y ago
You can probably get 80% of the stress reduction from 20% circularity.
- Gibbon1 7y ago
  Definitely that. Also for a cylinder under pressure the hoop stress is twice the longitudinal stress. Which says to me that vertically oriented rectangular windows are optimal. Probably even more so since the windows are aligned.

gmueckl 7y ago

Stress concentration, especially at sharp corners, is a well known phenomenon. During the second world war, the US manufactured transport ships in an assembly line fashion. The were launching these ships at rates of more than one per week. A lot of them had the same design flaw: the cargo hatch corners were not properly rounded off. This made them starting points for cracks in the hull. Amd once a crack is started, it continues to extend relentlessly under stress. As the story goes, a few of these ships were lost because the cracks grew so long that the hulls broke and sank.

nutcracker46 7y ago

It is unfortunate that people lost in the Comet crashes could not have enjoyed rides on better engineered aircraft, but that is how advancement really happens. Look at the DC-10 and B-737 Max. As to secondary walls in the pressure vessel, bladders, etc., take a look at actual designs for both windows and cabin pressure outflow valves are designed. They work well and can withstand quite a bit of adverse conditions. Those outflow valves are complex - usually consisting of motorized doors plus one or two spring loaded check valves for overpressure events. Usually there is another outflow valve for purging smoke or creating extra cooling flow for electronics under the flight deck. Going back to windows and other cutouts, there just wasn't enough sharing of expertise among ship and aircraft builders. There wasn't deep enough thought about effects of pressurization cycles. Now we know.

jjakque 7y ago

Real Engineering made a video on this exact topic, explaining the essence in less than 2 minutes: https://www.youtube.com/watch?v=7rXGRPMD-GQ https://www.youtube.com/watch?v=7rXGRPMD-GQ He went through bit more details said topic in another video "Why Are The Dreamliner's Windows So Big": https://www.youtube.com/watch?v=7-I20Ru9BwM https://www.youtube.com/watch?v=7-I20Ru9BwM

NamTaf 7y ago

Stress concentration factors like this are the bane of all structural engineering. Anything with too sharp an internal corner will concentrate stress and lead to material fatigue failure if there's any stress there. Given I've spent all morning staring at it in some finite element analysis work, I'm going to ramble about it to give myself a break. Sometimes dealing with it can be a pain, for example if you need two rectangular things to sit together along two faces, since you have to radius both corners with a larger radius on the smaller object. One way to combat this is to drill a hole in the corner, removing additional material but creating a little 270ish-degree-C-shaped radius in place of the sharp corner. Take what would've originally been a sharp corner, but then drill it out (example: [1]). What you lose on less material, you more than gain back with eliminating the stress concentration. This has the advantage that then a rectangular object can fit snug into the corner, because the smaller's sharp corner just sits inside the larger's C-shape cutout. The best visualisation tool of this phenomenon that I learned is force lines [2]. Where they bunch up closely together is where high stress occurs. It follows naturally that a sudden sharp corner will concentrate all those lines close to it, whereas a gentler radius encourages a more gradual transition. Think of the lines as similar to Isobars (probably the most well-known contour line) [3], in that the lines represent points of constant pressure/force, and so will sort-of repel adjacent lines (since there needs to be a smooth gradient between those lines). The more gradually they're forced to change, the more gradual that gradient between lines will be (represented by them being farther apart). Anyhow, all of this is also why you see countless concrete footpaths, driveways, etc. with cracks growing out from sharp corners in the concrete. Look at any internal corners where the concrete runs in an L shape. You'll not uncommonly find a crack growing from it. However, no one really cares (except for aesthetic properties) because it doesn't cause a functional problem, and forming concrete into straight sides is orders of magnitude easier than forming smooth curves. Bonus points if you notice that it's generally not parallel to one of the edges that form the corner, and can figure out why :) (I'd be guessing if I gave my answer, but I'm somewhat confident it'd be right: thermal expansion/contraction affects both sides simultaneously, meaning that the shear stress runs generally at 45 degrees to the corner rather than perpendicular to one, which would be the case if only one side expanded/contracted.) [1]: https://img3.gmdu.net/35780.0.jpg https://img3.gmdu.net/35780.0.jpg [2]: https://en.wikipedia.org/wiki/Force_lines https://en.wikipedia.org/wiki/Force_lines [3]: https://en.wikipedia.org/wiki/Contour_line#Barometric_pressure https://en.wikipedia.org/wiki/Contour_line#Barometric_pressu...

ganesh7 7y ago

You could also say the great innovators did not understand the technology they looted from the Germans. Or the similarities in design, e.g. integration of turbines in the wings, of the Comet and Nazi jet planes are merely coincidental I assume.

dTal 7y ago
The placement of the engines has nothing to do with metal fatigue from thousands of repeated 40,000 feet cruises with a pressurized cabin anyway; no German jet fighter was pressurized, so it's doubtful that they understood the technology either. Also, the only Nazi aircraft of any type with root-integrated jet engines is the obscure Horton 229 prototype - and apart from that single cosmetic similarity, it has absolutely nothing in common with the de Havilland Comet.
- ganesh7 7y ago
  Uhuh: "The Junkers Ju 49 was a German aircraft designed to investigate high-altitude flight and the techniques of cabin pressurization. It was the world's second working pressurized aircraft, following the Engineering Division USD-9A which first flew in the United States in 1921.[1] By 1935, it was flying regularly to around 12,500 m (41,000 ft)."
  - dTal 7y ago
    What on Earth is your argument? Everyone was experimenting with cabin pressurization - that wasn't even the first one. Are you seriously suggesting that that specific aircraft - a pre-war, single-engine piston craft - in any way influenced the design of the Comet? https://en.wikipedia.org/wiki/Cabin_pressurization#History https://en.wikipedia.org/wiki/Cabin_pressurization#History