Yep, hp is proportional to the cube of velocity. So, to get a 50% increase in velocity, you are going to need roughly 3x the power.

Well, not quite, don't forget that total drag = profile drag + induced drag.

True, profile drag that goes up by the cube of airspeed (profile drag hp that is), but on the normal side of the power curve the total drag curve increases less than the cube of airspeed. Recall that induced drag varies inversely with the square of airspeed (induced drag force) and varies inversely with airspeed (induced drag hp; power = force × speed). The total drag curve is the sum of profile and induced drag, and at moderate speeds the total drag hp curve is shallower than a cubic function, something closer to the square of airspeed (but mathematically different).

It's only at very high speed, when induced drag is almost negligible, that the total drag curve approaches the cube of velocity. And for the Raptor, at planned or at its apparently achievable speeds, induced drag is always going to be an appreciable chunk of total drag.

Well, not quite, don't forget that total drag = profile drag + induced drag.

True, profile drag that goes up by the cube of airspeed (profile drag hp that is), but on the normal side of the power curve the total drag curve increases less than the cube of airspeed. Recall that induced drag varies inversely with the square of airspeed (induced drag force) and varies inversely with airspeed (induced drag hp; power = force × speed). The total drag curve is the sum of profile and induced drag, and at moderate speeds the total drag hp curve is shallower than a cubic function, something closer to the square of airspeed (but mathematically different).

It's only at very high speed, when induced drag is almost negligible, that the total drag curve approaches the cube of velocity. And for the Raptor, at planned or at its apparently achievable speeds, induced drag is always going to be an appreciable chunk of total drag.

CFD is not required to determine the probable speed(s) vs HP. The rules of thumb that have existed for decades can reasonably predict those performance curves. Peter believes that the rules of physics do not apply to his work. Several have attempted to politely point out his mistakes, but he doubles down and trundles on.

Modeling and computer simulations whether FEA, CFD, computational chemistry (etc) are not as deterministic as you apparently expect. Engineering assumptions and mesh’ing changes can both have unanticipated and out-sized effects ... and the (cough) answers may be conflicting/confounding. Resolving such matters requires considerable experience and insight. Lather. Rinse. Repeat.

I'll try

FEA stands for Finite Element Analysis

The "finite element" is literal. You divide the part up into thousands or millions of tiny pieces and the analysis is run by looking at the interactions of forces & deflections between the pieces.

Step one is to create all the pieces. That is referred to as the "mesh". Creating the mesh on a complex shape requires an in-depth knowledge of how the modelling software is going to use the mesh.

At the ends of the mesh you have the boundary conditions. For example, where the part is bolted to another part. But that other part may have it's own deflections or vibrations under operation so those need to be considered in the analysis of the part in question.

You can run a model and get some beautiful looking results that are nonsense if the mesh and/or boundaries are done improperly.

And sometimes one style of mesh won't work for the whole part - you might have a discontinuity in the mesh around an area of interest so you might have to analyze the same part meshed multiple ways to get the whole picture.

It takes a lot of experience to get to go this right and people skilled in it are always in demand - but unfortunately it also tends to be something of a dead end career path. The organization tends to not let them go into another role once they get good at this.

here is an example of a mesh:

Tony, you're completely ignoring induced drag and by doing that you let my point go right over your head.

Napkin math for the Raptor predicts about 100lbf of induced drag at around 120 knots plus about the same amount of profile drag, for a ballpark horsepower requirement of about 75hp*. That's using the advertised weight and wingspan, applying a reasonable estimate for Cdi, and applying somewhat optimistic values for Cdp and/or the airplane's equivalent flat plate area**. At 180 knots the profile drag will probably be at least 200-250lbf while the induced drag should drop to perhaps 40-50lbf, for a total drag of about 250-300lbf, up from 200, which translates to about 135-165hp* required.

See the difference? That's not triple the horsepower to go 50% faster, it's roughly double; the difference is significant.

Push this airplane faster to 200 knots and certainly the power required will begin rise sharply.

The drag force at his 300 KTAS/FL250 design point is virtually the same as it is at 300 KIAS and sea level, although the power requirement up there is 50% greater, so about 200-250hp*... on paper.

Speaking of "on paper," as I mentioned, this napkin math is based on some optimistic assumptions, which brings us back to the current state of affairs of the prototype being draggier than he was hoping for.

And more about napkin math and paper airplanes, if he doesn't resolve his cooling problems to be able to fly for more than a few minutes and with the gear retracted, then all this math will be only on paper.


I hope all that analysis makes sense. My point is that you're oversimplifying it if all you say is "drag squared, power cubed."
Peter's done some overly simplified technical analysis... let's try to avoid doing that ourselves.






* 75hp will mean at least 100hp from the engine when accounting for gearbox losses and correcting propeller efficiency; 135-165 using the same estimates translates to 180-220, and that 300 KTAS/FL250 cruise point suggests well over 300hp from the engine.

** cleaned up, gear retracted, hopefully some refinements like replacing some un-aerodynamic antennas I can see in the videos, detailed attention to control surface gaps, fairings, etc.

Yes, but who would want to do anything else?

Tony, you're completely ignoring induced drag and by doing that you let my point go right over your head.

Napkin math for the Raptor predicts about 100lbf of induced drag at around 120 knots plus about the same amount of profile drag, for a ballpark horsepower requirement of about 75hp*. That's using the advertised weight and wingspan, applying a reasonable estimate for Cdi, and applying somewhat optimistic values for Cdp and/or the airplane's equivalent flat plate area**. At 180 knots the profile drag will probably be at least 200-250lbf while the induced drag should drop to perhaps 40-50lbf, for a total drag of about 250-300lbf, up from 200, which translates to about 135-165hp* required.

See the difference? That's not triple the horsepower to go 50% faster, it's roughly double; the difference is significant.

Push this airplane faster to 200 knots and certainly the power required will begin rise sharply.

The drag force at his 300 KTAS/FL250 design point is virtually the same as it is at 300 KIAS and sea level, although the power requirement up there is 50% greater, so about 200-250hp*... on paper.

Speaking of "on paper," as I mentioned, this napkin math is based on some optimistic assumptions, which brings us back to the current state of affairs of the prototype being draggier than he was hoping for.

And more about napkin math and paper airplanes, if he doesn't resolve his cooling problems to be able to fly for more than a few minutes and with the gear retracted, then all this math will be only on paper.


I hope all that analysis makes sense. My point is that you're oversimplifying it if all you say is "drag squared, power cubed."
Peter's done some overly simplified technical analysis... let's try to avoid doing that ourselves.






* 75hp will mean at least 100hp from the engine when accounting for gearbox losses and correcting propeller efficiency; 135-165 using the same estimates translates to 180-220, and that 300 KTAS/FL250 cruise point suggests well over 300hp from the engine.

** cleaned up, gear retracted, hopefully some refinements like replacing some un-aerodynamic antennas I can see in the videos, detailed attention to control surface gaps, fairings, etc.

Yeah, and we can infer some numbers by playing around with some basic performance equations, either numbers that seem achievable or "what if" numbers needed to achieve certain desired performance.

Your point about body lift is a good one- it is probably making some lift, which could have an overall positive effect on performance or maybe slightly worse. It's hard to say which without wind tunnel or flight test data.

Computerized analysis (CFD) is not so simple and not the be-all and end-all either, as a few guys have explained. It's one thing to model an airfoil section on a computer or one part of the airplane- say, a wing root or a fairing, something specific but not overly complicated. The whole airplane is a collection of too many parts for the current state of the art CFD to yield "good" data with certainty. That means your best guess is as good as mine, which means whatever seems reasonable for lift and drag coefficients- namely induced drag about 5-10% worse than theoretical "ideal" elliptical spanwise lift distribution, and approximating profile drag simply with x square feet equivalent flat plate area. Much earlier in the thread I mentioned flat plate area and that the horsepower-performance figures pointed to about 2 sq.ft., which would be a very, ahem, remarkable achievement for a five seater.

I am no engineer but I can certainly recognize that his cooling inlet is “unconventional” to put it nicely. I am sure that most inlets are of pitot style design like the Cessna Suck & Blow and P51 for a reason. It actually looks like it could be a low pressure point. I wonder if the flow could be burbling over causing turbulence affecting the upper half of the prop causing the pitch gyrations.

The cooling problem is that he's not ducting the cooling air through his radiators or oil coolers. The cooling air just enters the cowling, and ultimately flows out of the cowling, without being forced through a radiator or oil cooler.

Adding 50 pounds of water to his cooling system is dumb as a box of rocks. He could add 5 pounds of ductwork and get far more benefit, but he's addicted to solutions that don't fix anything (the 4 pounds of weight in the wingtip is a good example).

IIRC, the cooler for the induction air is mounted above the radiator. His bass ackwards, staged turbos are generating copious amounts of heat. The cooler is likely seeing inlet temps well over 200°F. (We’d know for sure if he had bothered to installed useful instrumentation instead of worrying about cup holders and air conditioning. ) Even if all the cooling air is forced through the radiator, it’s no longer cool when it get there. The tune, turbos and cooling are so poorly thought out and integrated together no band-aids are going to make a substantial improvement.

I tried to go to his website and look at the cooling system arrangement but didn't find much, and didn't want to go through all the videos. One place it still showed a top mounted variable scoop for a PT6 inlet.

new video out today. You will all probably be very surprised but the angle iron in front of his wheel wells did not solve the problem. The good news is he is making even more power now.


https://www.youtube.com/watch?v=7K7UjOnWcRU

I think he said he didn't run full RPM on his first flight, and one of the tests in this video was to see if he got more power at higher RPM. What engine does not provide more power at higher RPM??

Then he compared acceleration against a Cirrus to conclude he probably had 300HP. I thought he was talking about 350-400hp? I noted that in the comparison test, the Cirrus was at 1000' higher elevation.