The real savings come when this tech can be built into the airframe. The airframe can only bend and flex a certain number of times before it cracks and fails.
Todays airframes have a lifespan measured in flight hours, and an estimation is done as to how much turbulence will be hit per hour of flight. At the lifespan limit, the plane is typically scrap.
If this tech can reduce the flexing of the airframe during flight by 80%, you can probably get 5x the flight hours from the airframe before it becomes too weak to be safe (or more - half the flexing typically more than doubles the lifespan)
Alternatively, you can make the airframe thinner and lighter for the same number of operating hours (and that's what is likely to happen, since aircraft manufacturers don't want to put themselves out of business). Thinner and lighter airframe saves fuel and makes the aircraft cheaper.
First, I don’t understand what “build it into the airframe” means. These are sensors that are used to predict turbulence and then generate forces in opposition. Are you imagining small sci-fi rapid thrusters throughout that can generate enough thrust to counteract?
Secondly, the turbulence prediction is extremely hard for an airliner because it’s traveling so fast - you’d need sensors extended comically far forward and at that point you’ve got real risk of them breaking off mid flight meaning you would have to add significant amounts of weight to strengthen them (assuming you could). You’ve also got the problem that you need to retract this stuff on landing probably which adds more weight and complexity. Adding even more problems, generating sudden thrusts to counteract turbulence for a commercial airliner seems really difficult since that’s not how the engines work (eg you probably can’t generate a countervailing force quickly enough).
This is a neat concept but keep in mind this is a PoC on a very lightweight craft going relatively slow. It’s not clear how big/fast an aircraft it can scale up to. An easier turbulence reduction would be to mount the passenger area in something that could actively mechanically stabilize like optical image sensors. If you could decouple things so that the airframe could be repaired/replaced cheaply independent you everything else in the aircraft, that would be much more cost effective. However I suspect the mechanical stabilization itself would add a lot of weight/also need replacement and you wouldn’t see fuel savings I think, just a more comfortable ride.
These problems are hard, but have already been solved. The B-1 Lancer has active turbulence reduction built into the airframe, via the small canards on its nose. https://ntrs.nasa.gov/citations/19840005129
It's designed to work at Mach 0.85, and was meant to increase the lifespan of the airframe during low-altitude penetration flights where lots of turbulence could be expected.
Do you have a feel for how the maneuverability of a B-1 being controlled by the canards compares to that of a typical airliner?
It seems like a system like this would need to respond very quickly to changes in the air mass, and the weight and slow response of an airliner might make this system less feasible unless you could somehow measure airflow a reasonable distance in front of the plane.
I think the real idea is to have surfaces or controls distributed that could unload turbulence from individual surfaces nearly instantly.
This could be a few extra control surfaces, like a canard or actuated slat, or it could be through something like Active Flow Control -- https://www.scientificamerican.com/article/the-next-darpa-x-... -- where puffs of bleed air or electrostatics adjusts airflow rapidly.
I think systems-- like the B-1 Lancer highlighted by a sibling comment-- show it's not completely impractical for larger aircraft.
Neat. The best way to learn something is to post something wrong on a forum and wait for experts to correct you :). I hadn’t considered using puffs of air. I do still think adapting it for a commercial airliner may be tricky but it would be neat if one day a bunch of tech came together to realize this.
You don't need thrusters (which would be impractical). I think you can get most of the benefit by sensing with tiny canards in the front next to pitot tubes, and using electric motors to adjust ailerons/flaperons/spoilerons multiple times a second.
I think those are great examples of platforms that are not practical - they are only used in situations where they absolutely must get a plane to take off/land vertically, no matter the cost (in maintenance overhead or efficiency). I don't think they clear the bar in commercial aviation - a bleed air thruster system, even confined to the wing, would add a ton of complexity and hazards and weight. By comparison, the actuators, control surfaces, and airflow sensors are already there in commercial planes - in some planes it might even be possible to retrofit without introducing any new hardware, aside from more, bigger, more sensitive AoA sensors perhaps (which Boeing needs to do anyway, zing.)
I believe the final solutions for airliners would be using LIDAR to detect airflow in front of the plane based on the particles in the air. There's a few scientific articles about it if you search, and I believe from second hand information that the big manufacturers (airbus and Boeing) are testing prototypes for deployment on next gen airliners.
This may be equally scifi, but for sensors I'm imagining small jet powered drones flying in formation a few hundred metres ahead of the passenger plane.
Given how slow and cautious the industry is (and for good reason), even if someone has been working on this for a decade already, I don't expect to see anything like that before 2030.
Yeah unfortunately I don’t think such tech exists and at our current tech level I wouldn’t imagine it would work and could present real safety issues - you’d have fuel pipelines running throughout the airframe which is likely a significant amount of really serious fire risk, all of this adds a significant amount of weight and maintenance cost, I don’t believe such adjustment thrusters exist but I could be wrong, and I’m sure you would still get things wrong in your prediction which means your thrusters would add wear and tear on the airframe as well.
This article is about a PoC where they put sensors extended out of the airframe of a small slow moving aircraft and demonstrated a 60% prediction accuracy for a 10% fuel savings. It’s unlikely this approach would scale up to an airliner. The sensor problem might be but I have big questions about the adjustments an airliner moving at 600-900 mph can make to successfully counteract the prediction.
Doppler radar systems can detect airflow vectors in arbitrarily large areas, so I imagine that we have the tech for that portion, but I too am concerned about trying to take something with that large a mass and surface area and counteract turbulent forces with it. That's also going to stress the airframe and avionics, cause maintenance schedule changes, etc...
Yeah, i was joking. Those thrusters would have to be insanely powerful to accelerate such a big mass fast enough to counteract the rapidly changing forces, caused by turbulence.
> airframes have a lifespan measured in flight hours
For general aviation, yes. For pressurised cabins, life is measured in pressure cycles. A long flight wears the structure about as much as a short one; it's the inflating and deflating that counts.
This. My understanding is that airliners basically always hit the pressurization cycle limit before anything else, so wing turbulence cycles are not important.
At least with commercial aviation, unless you can’t get ahold of new planes they get replaced much earlier than that due to wanting ever more efficient planes and more stringent noise regulations.
The old clunkers still around are mostly used in sanctioned countries, or where a market failure has failed to provide a decent equivalent (757/767)
> or where a market failure has failed to provide a decent equivalent
A whole lot of weird, or just simply old planes still operate in Canada for this reason. Until this year you could still find 737-200 flights in the north.
> “ Todays airframes have a lifespan measured in flight hours”
Isn’t the lifespan related to the number of cabin pressurizations (not flight hours).
Which is why long haul planes like 787 have longer lifespans than a 737, because 737 are doing way more quick turn trips (more pressurizations) than a long haul international flight plane.
I am not aware that the aiframe is the main reason why planes are taken out of service. Fuel inefficiency is typically the primary reason, plus increased maintenance cost. I can't remember an accident of a major airliner as a result of the airframe failing post the 1970s.
Will be interesting to look at fatigue of the control surfaces though, if they get used many times more per flight to adjust for minor turbulences.
> I am not aware that the aiframe is the main reason why planes are taken out of service
This is true--- these days. We've been tending to replace planes before the airframe life limits are met.
> I can't remember an accident of a major airliner as a result of the airframe failing post the 1970s.
Well, on the other hand, this is mostly true because airframe life limits and expensive inspection programs that we established after horrific accidents.
> Will be interesting to look at fatigue of the control surfaces though, if they get used many times more per flight to adjust for minor turbulences.
Control surfaces are on bearings and have, in general, a lot of excess strength to minimize internal deflection. Even if there were a trade in wearing out control surfaces faster, they are a more easily inspected and replaced part of the airplane.
I think route optimization that this will enable (not avoiding certain turbulence) might be valuable (save fuel) - I'm very excited about the potential implications per flight that this might have
The lighter airframe is really interesting, too, given the ongoing cost savings
Whenever I'm flying on the B787, I prefer to sit right behind the wing to watch the flaperons do their thing. I know it's a little different than this, as it's more of a PID driven behavior rather than an active predictive system. But regardless, it's quite impressive to watch how they are constantly reacting despite the plane seemingly not moving at all. They move very delicate and precise, which must be difficult to do when travelling at 500mph.
>as it's more of a PID driven behavior rather than an active predictive system
Self tuning and active predictive PID controllers are also a thing aided by the 'ye olde' faithful Kalman-Filter. At least I remember reading about them in research papers.
Now what exactly from those has materialized in commercial applications, I have no idea, since it's not like they publish such in depth info in the public facing spec sheet.
> on long pole masts that placed them some 2.65 m (8.69 ft) forward of the leading edges. At cruise speed, that's enough to give the system a tenth of a second's worth of advance warning before turbulence hits
2.65m */ 0.1 seconds = 60 mph.
Airliners fly about 500 mph, so something about this math is far off...
> The company says it'll have a system commercially available for light aircraft in 2024. It's looking into a version for eVTOL air taxis by 2026, and hoping to have a system relevant to commercial airliners by 2030. Godspeed, team, the world's airline passengers – not to mention cleaning crews – need you to succeed.
They are not targeting airliners for the current generation.
>They are not targeting airliners for the current generation.
Like someone else said in the comments, (if true) this tech seems to have already been used in military bombers for a long time now, so to me it's weird it hasn't made it to civilian aircrafts already from the civilian arms of military contractors, and instead needs to be reinvented by a start-up.
Maybe they meant the cruise speed of the test aircraft. Some quick googling reveals it as a Colomban MC-30 Luciole [1], which according to Wikipedia [2] has a cruise speed of 110 mph and a maximum speed of 120 mph. Mounting the long instrument arm to the tiny aircraft probably doesn't do it any favors (seriously, the aircraft looks tiny next to a Cessna), so cruising at 60mph sounds reasonable.
The pole is on the nose, the time available to react would be before reaching the leading edge of the wing. You'll need to include the length of the front of the plane in question. At least as I understand it.
There are real time industrial sorting systems that use sensors and air puffers to sort items off a conveyor belt 30 ms after they pass by the sensor array.
So the class of technology they need to make this works already exists in production. So the real problems will be how they’re sensing, and can they make them survive hundreds of thousands of air miles.
I’ve seen videos of those. It’s honestly mind bending.
A shower of something like nuts or cherries halls a few feed past a bright area and the bad ones just magically sort themselves out into a different stream away from the main flow.
The last one I saw was using X-rays to sort aluminum. Apparently there’s a lead alloy out there. You don’t want to make window frames out of those. But making more of that alloy is fine.
I would guess that it gives them enough advance notice to predict turbulence, based on the delta between forces on the plane and forces on the poles. If there's a sharp gradient in forces, that gradient seems likely to increase, in a partially predictable direction.
Yeah I'm guessing these poles just get out in front of the wing aerodynamics enough that they can sense whats going on with the air so that air pockets/turbulence can be predicted. Either that or we are missing a decimal place, maybe they mean 10 milliseconds?
One test aircraft used 2.65m poles extending off the front of the wing. One test aircraft used a boom extending off the nose. The model airplane used a big rig sticking out front and teeing off to the sides.
The 787 is far more comfortable than earlier aircraft in regards to vibration, sound, air quality, and lighting - there's a lot of internal upgrades. My mother who has had severe headaches and sinus issues that are extremely sensitive to altitude pressure when flying through the 1990s-2010s has had no problems since flying on the 787.
The 787 cabin pressure is equivalent 5000 ft above sea level versus 8000 ft for older planes. Essentially Denver versus Aspen.
They also created a 3D microphone array that they used to map the sound inside the airplane and spend the dampening budget in the places that will get the most impact.
In recent years I've heard pilots announce that they are re-routing to avoid turbulence. I'm not sure what's new: the announcement of this, or the ability to do so.
Anecdotally, I fly round trip out of SEA ~3 times a year and experience very bad turbulence on about half the flights. Earlier this year it was bad enough to suspend drink service.
> Earlier this year it was bad enough to suspend drink service.
If drink service wasn't suspended on others, it wasn't very bad turbulence. A rule of thumb is that if your seat belt isn't hurting you, it's moderate or lower intensity.
About a decade or two ago, turbulence seemed worse. My uncle told of a time when he saw people hit the ceiling. I've rarely had issues, although plenty of smaller pockets where service does get suspended. I fly out of SEA, but in my opinion, DEN is much bumpier because of the sheer winds from the Rockies.
My rule of thumb is if the drinks didn't fly into the air and spill, then turbulence is minor.
The idea of looking ahead at turbulence and preparing for it is good, but it has to be done with a technology that doesn't require placing an enormous pole in the air, e.g. laser, radar, etc.
Yeah that’s one problem for airliners. The others are getting accurate predictions to happen far enough into the future that the contraction would work and being able to generate high levels of thrust into a specific direction at a moments notice to counteract it.
You need to be able to get the information fast enough to be able to make the control surfaces adjust in time.
The longer the plane, the further between the nose and the leading edge of the wing. So the more time you have.
also aren’t turbulent more likely when you are lower than at final cruise altitude? Wouldn’t the plane be going slower than anyway for fuel savings in the denser air?
Even if it would only work on longer/larger jets it could still be very helpful.
The front of the airplane is very far ahead of anything that could possibly act on that information. Not nearly as much in the case for this test plane.
That's a really good point. Napkin math on an Airbus A321 gives about 81 milliseconds of time to react, not far off from the 1/10th of a second they cite for their prototype.
The projected climate-driven increase in turbulence as a marketing point is interesting, but there are caveats. Research studies do point to turbulence increases in certain seasons and latitudes:
> "Climate modelling studies have indicated that the volume of airspace containing moderate-or-greater clear-air turbulence on transatlantic flight routes in winter will increase by 40%–170%, relative to pre-industrial times, when the CO2 is doubled"
The CO2 doubling point is expected to arrive in ~60 years at current fossil fuel combustion rates, but I'd expect by that time a very large fraction of short-distance air travel will have moved to (electrified) high-speed rail as it's far cheaper per distance traveled. Also, as others note, this technology doesn't seem applicable to trans-oceanic jet travel.
In 1951, Kraft issued NACA Technical Note 2416 that proposed a theoretical solution to the problem. Tests first on a modified DC-3 and later on a C-45 validated the theory, and by 1955 the system was perfected.
Remote control airplanes have been using gyros to achieve the same effect for some time now. It's a complete game changer and changes the amount of enjoyment in a huge way. Instead of telling the airplane what to do, you're more so telling the airplane where to go.
I'm not suggesting gyros are the correct way to do this. But I can say the difference is night and day.
It isn't a gyroscope that physically resists rotation. It is an active sensor that is used to control a PID loop to control the actuators and reduce the disturbance.
The technology in the linked article is even different, they put a sensor far in front of the wing and preemptively control the surfaces to counteract turbulence.
The gyros in RC are misleading. They are a sensor (gyrometer) which isn't using mass to counter movement, the are merely letting the plane know when it has shifted so it can move the control surfaces accordingly. Gyros have mostly been replaced by accelerometers at this point.
I've occasionally thought about getting into RC planes, but the piloting of them always seemed too hard. I'd love some links to what you're talking about if you have some to recommend?
I wonder if adding spinning wheels arranged in multiple axis might be simpler ways to counteract the forces. Might be too hard to do it in a safe/lightweight way and spin the wheels fast enough to counteract the forces at play.
To get a strong enough reaction wouldn’t you need very heavy wheels that the airlines would be unlikely to be interested in carrying due to extra fuel use?
I was thinking more of a constantly spinning wheel and conservation of angular momentum but maybe that doesn’t help if most of the motion of turbulence is translational? I’m sure it’s an impractical idea.
You're describing a Control Moment Gyroscope (CMG), famously used on the ISS. They have very high torque output per unit power and mass/volume. They and reaction wheels (fixed axis of rotation) are convenient for spacecraft applications because they don't require an external medium to push against. Aircraft are surrounded by air, so they can use control surfaces instead.
I wonder how this will effect the wear rate of moving parts on an aircraft. Normally they are "fixed" right? And now they are going to be moving constantly, during the entire flight.
The control surfaces? The pilot (or autopilot) is fairly constantly making small corrections, so no, they’re not ‘fixed’, but it’s true this is a much higher frequency movement than normal control.
That’s a good question. But someone else in the thread said that this might allow the wings to last longer if you can ensure that the forces on them are lower on average.
It’s certainly going to be easier to replace hydraulic cylinders or something for the control services than the entire wing (which I’m assuming no one does).
So it may still be a net win in maintenance terms.
Hmm, I was thinking small flywheel / gyro units in the plane fuselage itself similar to active roll damping in boats. I guess active control surfaces make more sense for airplanes, without knowing much about how turbulence affects felt vibration in the fuselage.
Fighting turbulence must have an impact on fuel efficiency and range. Additionally, pressure probes add weight and drag to the aircraft. Overall, I think these kind of solutions look brittle and error-prone.
Pilots still fight turbulence today, usually by changing altitude. If this system allows the plane to stay at optimal (but turbulent) altitude, it could save fuel.
Drag force increases with the square of speed, so it’s probably more to do with a longer flight costing more operational hours than a short one. I can’t sell flights on a plane that hasn’t landed yet.
I mean, they do anyway, but eventually it leads to refunds, and fines from the FAA. Actually on second thought it’s probably not physics or logistics, it’s the FAA.
was just wondering how pilots know ahead of time that there is turbulence a minute or 2 out (obviously can't see it in the distance) and it's the least technical way I thought possible
This is terrible. Turbulence is the only time on a flight where adults behave like adults [1]. Even their sphincters pucker preventing them from degassing.
"Nobody likes to fly through turbulence"
Thats not true! There are dozens of us! (Actually, quite a big more, I read a stat that we're about 5% of the fliers)
[1] kids be kids, of course - as a kid I once ran up and down the aisle during take off, so Im cool w/ kids.
My wife has a fear of flying but she likes some mild turbulence. It helps her sense that flying is a physically sensible thing, not some magical thing that shouldn't work.
That kind of makes sense, it's like how rock-climbing on a rope is scary until you actually fall (or descend) and can feel the tension of the rope actually supporting you.
Even on a gut level, feeling the plane bounce against the "road" gives you a confirmation that, yes, in fact, it is bouncing against something (even if that something is just air under the wings).
Yea, but a quick prayer on take off/landing is not unwarranted. Normalized by journey instead of mile, flying is more dangerous than driving (according to Wikipedia anyway).
That transport comparisons table puts it at 3x per trip. This is like praying for your safety because you’re going into work and stopping for groceries on the way home.
There have been five fatalities of US airline passengers in the last 10 years. That’s billions of passengers. It’s incredibly safe.
The 346 people who died in Boeing 737 Max aren't in the US' tally because of sampling error, not because those accidents could not have happened in the US.
"A total of 42,939 people died in motor vehicle crashes in 2021. These deaths occurred in 39,508 crashes involving 61,332 motor vehicles. This was a 10 percent increase in deaths compared with 2020."
When you say driving is also safe, it doesn't really seem like a reasonable comparison when 85,878 died in motor vehicle crashes for every 1 person that died from flying.
I think these need to be unpacked and the IIHS data is missing a couple metrics.
Right now the FAA is bragging about serving 2.9 million passengers a day. I had no idea it was that high. That’s almost 1% of the population. The IIHS death rate works out in the neighborhood of 1 in 8-10k. But I think that’s per driver, not per trip. So we are stuck on the per trip bit again.
Meanwhile 5 fatalities in 10 years (~3652 days) is less than 10.5 billion person trips (assuming passenger rates climb over time) which is around 1 in 2 billion. So unless we are living in a statistical fluke that’s about two orders of magnitude off.
It has a chart normalizing by trip. Someone posted this on HM a few days ago.
Also, the statistical difficulties of this, and therefore large error, are significant. I see at least two:
1. 5 fatalities in ten years in the US neglects the 346 fatalities of passengers who died in two 737 Max flights before being recalled - flight deaths that, being 100% Boeing's fault, aren't on the US' tally by stat fluke (lookin at you Southwest).
2. actually its not luck that the hundreds of 737 dead didn't end up in the US' tally. Airline deaths are rare enough that there are not enough flights to sample the death rate properly.
So you normalize car deaths by plane deaths and not by mile (plane wins) or journey (car wins). I appreciate your dedication to dimensionless numbers, but that's kinda weird.
There are orders of magnitude more driving than flying by mile and number of journeys.
I'm not sure how this reply relates to my post, which didn't mention prayer at all. You can pray for a good parking spot for all I care, it's really none of my business.
the counterpoint is that this is a system that would only be engaged at cruising altitudes, and which would probably be manually turned on. the problem with mcas (aside from using a single sensor for flight control) was that it had large effects at low speed and problems near the ground don't give you time to react
The issue with the MCAS-fiasco was that it (by design) was not communicated to the pilots, so when it failed, they didn't know immediately what to do. If they had known, they would have cut out the trim motor and left it off and trimmed manually until landing. In the accident flight, the crew repeatedly turned on and off the electric trim and never figured out why it started moving the trim every time they turned it on again (though, yeah, it's an open question why they did this more than one cycle....)
The second accident flight shows knowledge of the system wasn't enough. On that flight, they did turn off electric trim, but weren't able to manually trim because the force required to manually trim from the limit was too much. So they turned electric trim back on, and shortly after they stopped touching the electric trim inputs, MCAS resumed setting the wrong trim.
IMHO, the second accident had a better chance of survival, but it wasn't enough. IIRC, the flight before the first accident also had erroneous MCAS activation, and the flight crew did turn off electric trim and did it manual, but it wasn't treated as a must fix maintenance item, because MCAS was hidden.
There really should have been a separate shutoff for MCAS apart from the electric trim switches. Limited activation authority will hopefully be sufficient, but doesn't satisfy my airchair aerospace engineering demands.
counterpoint: the primary problem wasn't that mcas was hard to turn off well. it was that it was a critical system without any redundancy and therefore failed about 100x too often. if mcas had been based off of 3 sensors like it obviously needed to be it would have failed (and known that it failed so it could turn itself off instead of pitching down the plane) one a decade or so rather than a few times a year.
Todays airframes have a lifespan measured in flight hours, and an estimation is done as to how much turbulence will be hit per hour of flight. At the lifespan limit, the plane is typically scrap.
If this tech can reduce the flexing of the airframe during flight by 80%, you can probably get 5x the flight hours from the airframe before it becomes too weak to be safe (or more - half the flexing typically more than doubles the lifespan)
Alternatively, you can make the airframe thinner and lighter for the same number of operating hours (and that's what is likely to happen, since aircraft manufacturers don't want to put themselves out of business). Thinner and lighter airframe saves fuel and makes the aircraft cheaper.