# Can wing-tip vortices be reduced/eliminated with a rear-facing propeller near the wing-tip?

Let's say we have a propeller-driven aircraft in the pusher configuration (rear-facing propellers). If we could somehow put an engine at the wing tip (ignoring structural concerns), and maybe make the propeller rotate in the opposite direction of the way the wing-tip vortex wants to spin, would this eliminate or at least drastically reduce the wing-tip vortex?

• It's been tried before.
– fooot
Jan 22, 2016 at 4:05
• @fooot a fascinating aircraft that seems to imply the answer is yes. However I envisioned the propellers to be a little proportionally smaller and the wing shape to be much more conventional. Hopefully the vortex-cancelling effect will be the same. Jan 22, 2016 at 4:29
• Vmca on that aircraft had to be pretty high! Jan 22, 2016 at 6:22
• @RalphJ I think the idea was specifically to have more lift and control at low speeds by using the propellers to boost the wing efficiency. The article suggests that was actually achieved. Jan 22, 2016 at 7:05
• That might actually be an interesting design to revive for hobbyists. It looks to me that it was designed to optimize space usage (hangar and deck) and robustness for carrier operations. I think there are people who'd like to have a compact and robust design. Although I kind of wonder what happens if you have engine or propeller problems... Jan 22, 2016 at 7:10

Yes. From an aerodynamic standpoint it makes some sense to use the wingtip vortex in combination with a propeller.

In this article by Sinnige et al., the researchers modeled and tested a propeller in various positions along a semispan, concluding that there were real gains to be had due to the increased span efficiency. Up to 15% less drag was measured when the propeller was positioned at the wingtip compared to a conventional mounting:

By positioning the propeller at the tip of the wing, the slipstream interacts with the flow around the wingtip, thus affecting the roll-up and downstream behavior of the wingtip vortex. PIV measurements downstream of a propeller–wing model showed that this leads to a reduction in overall swirl with inboard-up rotation, for which the swirl in the slipstream is opposite to that associated with the wingtip vortex. At the same time, the system performance was found to improve due to a reduction of the wing induced drag, leading to the conclusion that the decrease in swirl causes a reduction in downwash experienced by the wing.

Apart from the change in drag, the interaction of the wing with the propeller slipstream also modifies the wing lift. The locally enhanced dynamic pressure increases the lift over the spanwise part of the wing washed by the slipstream, which is amplified by the induced swirl for the case with inboard-up rotation. As a result, a strong spanwise variation in lift occurs with the propeller on. The induced velocities caused by this lift gradient lead to a spanwise shearing of the slipstream. With outboard-up rotation, the swirl in the slipstream acts to locally oppose the increase in wing lift due to the propeller-induced dynamic-pressure rise. Compared to the inboard-up rotation case, this leads to a reduction in wing lift at a given angle of attack, thus also a reduction in maximum lift coefficient. Furthermore, the direction of the spanwise shearing of the propeller slipstream is inverted on both sides of the wing.

To quantify the potential aerodynamic benefits of the wingtip-mounted configuration, a direct comparison was made with a conventional configuration, with the propeller mounted on the inboard part of the wing. The increase in wing lift due to the interaction with the propeller was 1 - 4% smaller for the wingtip-mounted configuration than for the conventional configuration. For the latter, the enhanced dynamic pressure and swirl in the slipstream acts over double the spanwise extent, and on a part of the wing where the section lift is higher than for the wingtip-mounted configuration. At higher angles of attack, the lift advantage for the conventional configuration could be further enhanced by the local angle-of-attack increase in proximity of both sides of the nacelle.

In terms of drag performance, on the other hand, the wingtip-mounted configuration showed superior performance. At a wing lift coefficient of $$C_L = 0.5$$ and a thrust coefficient of $$0.09 < C_T < 0.13$$, the drag reduction amounted to about 15 - 40 counts (5 - 15%) compared to the conventional configuration. The aerodynamic benefit of the wingtip-mounted configuration further increases with increasing wing lift coefficient and propeller thrust coefficient, leading to a drag reduction of 100 - 170 counts (25 - 50%) at $$C_L = 0.7$$ and $$0.14 < C_T < 0.17$$. An analysis of the wing performance confirmed that this drag benefit is mostly due to a reduction of the wing-induced drag. Compared to the conventional configuration, a relative increase in span efficiency of up to 40% was measured for the wingtip-mounted configuration. Although the exact drag benefit will be specific to vehicle design and operating conditions, it is concluded that the interaction between the propeller slipstream and the wingtip vortex leads to a clear drag reduction for the wingtip-mounted configuration.

Keep in mind this is purely aerodynamical, of course, and from a structural standpoint things might not be so clear-cut. While weight in the wings helps alleviate the bending stresses generated by lift, a large mass at the tip can lower the first bending eigenfrequency to unacceptable levels, risking a coupling of the wing oscillations with some other mode in flight or on the ground.

As far as I am aware wingtip fuel tanks have mostly been used on relatively low aspect ratio wings, which lends some credence to the magnitude of this problem.

• Well done for spotting that paper! Also, putting the propeller before the wing makes a ton of sense. An additional problem with having this weight at the wing tip: Landing. Passenger aircraft must be able to land at constant 3° glide slope (in case the pilot doesn't pull up in time) -- with nice slender wings, that means unless your engine is very light, you have a problem. So I would still not expect to see this in practice, at least not without a host of other smart ideas to make it work.
– Zak
Aug 13, 2019 at 1:05
• That paper had been sitting in my "to read" pile for half a year, this was more of a happy coincidence. Also, good point about wingtip clearance; there is also the issue of FOD damage for engines hanging outside the paved width of a runway. Aug 13, 2019 at 14:37

Eliminate, no. The wingtip vortices are inherent part of lift generation. There is no lift without wingtip vortices. The wake vortices are carrying the momentum that was given to the air to produce the lift, and to cancel them, you'd have to give the air the opposite momentum, which would negate the produced lift. See How does an aircraft form wake turbulence?

Improve efficiency, yes, a little bit. The propellers would cause downwash in their span outside of the wing tips, effectively extending the wing span and longer wing span means, slightly, less induced drag.

However, such design would have extremely poor single-engine handling, not only because the thrust would become very asymmetric, but also because the lift would partially depend on the engines and therefore the side with stopped engine would also loose some lift, and deflecting aileron to compensate would generate more drag to make the thrust even more asymmetric. And it would have poor all-engine-out performance too due to the loss of lift. Does not sound like optimal approach when similar benefits can be obtained without all these problems by using longer wings.

• "The wingtip vortices is what generates the lift." Wrong, wingtip vortices are a by-product of lift generated by the wings, but they generate drag by themselves. Otherwise why add winglets? Aug 5, 2019 at 21:33
• I agree with Micheal Hall. Otherwise, why not put some vortex generators on the wingtip to increase vorticity? Would putting propellers behind the wingtips to cancel the vortex out destroy the lift? The wingtip vortices are a side effect of having a lifting vortex in your wing (precisly: spanwise change of strength of the lifting vortex) -- which itself is as much a mathematical construct than an actual thing.
– Zak
Aug 5, 2019 at 23:22
• Correct: single-engine handling would be horrible (and also putting such masses at the wingtips is a recipe for aerolastic catastrophes), and ailerons interacting with propellers must be a nightmare in terms of stability and control (and the propeller, too) -- but the propellers wouldn't generate very much lift. Also: making the wings longer isn't always an option.
– Zak
Aug 5, 2019 at 23:27
• @JanHudec You are correct in saying lift cannot exist without vortices. Michael and Zak are also correct in saying wingtip vortices do not contribute to additional lift. Wingtip vortices are what contribute to the induced drag, a finite-span 3D phenomenon. In infinite span, induced drag would disappear and you would get the 2D result, and you will still have bound vortices.
– JZYL
Aug 7, 2019 at 12:47
• This thread is becoming ricidulously long. I suggest you look at @AEhere's answer and the linked paper there: Fig 24 nicely shows how lift is almost unaffacted by moving the prop to the wingtip, but drag is nicely reduced. There's more figures showing how the vortex is weaker if the prop is turned on. I'm tapping out now.
– Zak
Aug 13, 2019 at 1:10

Doug MacLean has a great synopsis of wingtip devices that is worth reading, and gives a good intuition as to why propellers could reduce, but likely not eliminate, the wingtip vortex.

To summarize: the wingtip vortex (really, the rolled-up vortex wake sheet), creates a downwash at the wing which effectively reduces the angle of attack of the wing relative to the freestream. This downwash increases with the strength of the wingtip vortex, which is proportional to the lift of the wing. The induced downwash angle $$\alpha_i$$ can be compactly given as $$\alpha_i = \frac{C_L}{\pi e AR}$$.

To compensate for this downwash, the plane has to fly at a higher angle of attack to achieve some amount of lift. A component of the lift ($$C_L\sin\alpha_i$$) is now pointing in the streamwise direction. This is induced drag, and if the $$\alpha_i$$ is small you get the conventional induced drag expression $$C_{D_i} = C_L\sin\alpha_i = \frac{C_L^2}{\pi e AR}$$.

So how does this change if we add another source of vorticity? It won't reduce the lift of the wing; the vorticity generated by lifting surface will still be present. To remove the induced drag ideally our new source of vorticity will have opposite sign and equal magnitude, so as to create a corresponding upwash at the wing, which would rotate the lift vector back towards the freestream direction.

Can the propeller do that? It's worth thinking about where the vorticity (swirl) comes from in the propeller wake. Swirl is the tangential velocity imparted to the wake; it's essentially due to viscous losses in the blades. Well-designed propellers try to, among other things, minimize this type of loss. If you had a propeller well-designed to provide forward thrust, you would need to put in a huge amount of energy that the viscous losses counteract the entirety of the wingtip vortex; likely much more than you need for forward thrust. In a typical jetliner wing, the lifting force in cruise is something like 20x the propulsive force. Swirl losses, like tip vorticity, are proportional to the thrust produced by the propeller, which is an order of magnitude smaller than the wing lift.

In theory you could design a propeller to add more swirl to the wake, counteracting the tip vortex more efficiently. But this would reduce the efficiency of the propeller at creating forward thrust, so is unlikely to give a net benefit.

As the testing shows there is some benefit to wingtip mounting. You can't design a propeller with no swirl, so you might as well get some credit for it by putting it at the wingtips (if you can accept the weight and OEI tradeoffs associated with it). But it's not going to entirely (or even substantially) eliminate induced drag.

TL:DR Propeller swirl and wingtip vorticies are both due to inefficiencies in the two systems. These inefficiencies can cancel to some extent, but they are both proportional to the amount of force being produced by the wing/propeller. Since the wing is producing ~20x the force of the propeller in cruise, the effect of its inefficiency will dominate.

• Welcome to Av.SE! Aug 7, 2019 at 18:46

The wingtip vortices carry some energy, and not leaving it behind seems like a good idea. That's why winglets are a thing, after all.

So, what if you put a propeller at the wingtip, aligning the axis with the vortex core? The propeller would "see" the incoming vorticity, and the blades would get accordingly larger local incidences, thus generate more forwards force -- a little like putting guide vanes in front of it. Alternatively, the propeller could rotate a little slower or reduce blade incidence a little to get thrust back to where it was. The swirl which the propeller produces would be accordingly reduced, so there'd be overall less vorticity and swirl in the flow behind the aircraft.

That all sounds pretty good so far.

However, there are a few drawbacks, from least to most severe:

1: The wintip vortex is pretty strong at its core, but angular velocity reduces quickly as you move away -- this means that the innermost section of your propeller gains most additional incidence, but since it's also moving slowest and the blades are thickest, it doesn't produce a lot of thrust anyway. The outer bits won't see much of an effect, as their own circumferential speed will be much higher than the vortex at that position

2: When mounting the propeller directly behind the wing, the blades will pass through the wing profile wake on the inboard side, where airflow is significantly slower. In the worst case, if the flow on the wing separates, propeller blades could go through a significant zone of "dead water", which means less thrust, and more mechanical load on the blades. Also more noise. Most existing pusher configurations have the propeller mounted at some distance from the wing to reduce this effect. But if you did that with wingtip pusher propellers, it would just make reinforce an even worse problem...

3: The wingtip vortex doesn't align nicely with the trailing edge of the wingtip, let alone the propeller axis. Depending on flight condition, the vortex will be stronger or weaker, and at large incidence, it becomes more of a smeared-out vortex sheet -- imagine lots of small vortices released from points along the outer wing edge, propagating in downstream direction. This means at many flight conditions you'd have the vortex not really matching with the propeller, this diminishing the desired effect, but in some other conditions, you'd have a strong, well-focussed vortex hitting the propeller somewhere off-center, and having your propeller blades rotate through that causes bad vibrations, and might also separations on the blades, which means lost thrust and horrible noise, coupled with either having to reinforce the entire drive train (or facing much higher wear and tear). You could make a propeller aerodynamically more robust to such things, but that will always come at the expense of efficiency, and that's what we set out to gain in the first place...

Pusher propellers are not very efficient to begin with (because putting the whole aircraft into the swirl coming from the propeller is still more efficient than exposing the propeller to the wake of the aircraft), and mostly used for stability reasons (This has to do with the pitching and yawing moment produced by a propeller in inclined flow) -- so although aligning a propeller with a vortex, in isolation does indeed make sense, there are too many real effects keep this from improving efficiency over a regular old front-mounted prop.

So, can't we do anything with that vortex? Oh yes, you can! You could place a little wing inside the upwash just outside the wingspan, also known as "increase the wingspan" -- more wing makes more lift, but the vortex does not get stronger. This is why sailplanes have such long, thin wings. Or, if you can't make the wing any longer (wing root bending moment too large, size restrictions...), add the little bit at an angle! The classical vertical winglet works like this: It redirects the inwards-headed flow above the wingtip to go straight downstream, and this produces mostly an inwards-facing force but also a forward facing component ==> this means it does exactly what the propeller can't do efficiently, which is weakening the vortex and deriving a little forwards force from it. These days, most winglets are some sort of blend between a wingspan extension and the classical winglet.

• How would a turbine engine placed on the wing tip to receive the tip vortex in its entirety?
– Muze
Aug 5, 2019 at 23:24
• @Muze: I typed "Turbofan vortex ingestion" into duckduckgo, and wouldn't you know it, this is the first result: ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19750018896.pdf -- somebody testing what happens when a wingtip vortex gets into an engine. It's not good. Stall and surge margin are reduced, and so is efficiency -- unless you hit exactly straight on, but you virtually cannot. Ground vortex ingestion is also a longstanding problem for turbofan engines: aviation.stackexchange.com/questions/21219/…
– Zak
Aug 5, 2019 at 23:35
• "Pusher propellers are not very efficient to begin with" – are you sure? Most evidence points to the contrary. Aug 7, 2019 at 18:42
• @Peter -- I have no proper literature to hand right now, but the fact that very few aircraft use pusher props (and that most pusher props are seen in photos from the Age of Experimentation) does point that way. The way I learned it: Pusher props avoid penalties for having the fuselage sit in the accelerated, swirling air from the prop, but degrade prop performance, which is (usually) worse. They're also audibly louder (can confirm from personal experience). If they are used, it's mostly due to stability reasons or to clean up the flow on the rear. Reality is likeley more complex than that ...
– Zak
Aug 8, 2019 at 14:08

You can't stop them from forming, but in the case of rotor blades you can reduce the tip vortices to the point that they are almost gone by the time the next blade arrives.