DF inlet and ducting design
By: Bob Parks

Inlet area-- basically the bigger it is, the less critical the shape is.
You can run a very tiny inlet (50% of fan swept area), but you have to be VERY VERY VERY careful about inlet shape and how the duct cross section enlarges out to the fan diameter. Probably the best example of doing this, and doing it right is the BVM F-86 DF version. It could be a bit better if the inlet lip was fatter, but keeping it looking scale prevented that.

If you have to turn the duct (side inlets or whatever), then its a lot easier to keep good flow if the cross section area is decreasing as the flow goes along the duct. Turning and increasing the area is asking for trouble. Yes, it can be done, but plan on doing flow testing (a clear fiberglass duct with tufts can be useful for this). Note that flow testing never hurts regardless. You can design and analyze all you want, but the proof is in the testing.

When looking at duct cross section areas, dont just look at the top and side view, go plot cross sections (do them perpendicular to the duct centerline) and measure the areas. Most likely you will be unpleasantly surprised at how bad it is. Adjust the shapes until the area distributions look good AND the the shapes look "right". This is VERY critical on bifurcated inlets and exhausts. They can be astonishingly bad!

Having a nice, well rounded inlet lip is good, since it lets the inlet work efficiently over a wide range of flight speeds without too much drag or too much thrust loss. Using a 2:1 ellipse is a good starting point for the lip cross section. Having the lip wall thickness at least 15% of inlet diameter (or the diameter of an equivalent area round inlet) is also good.

For fixed wing DF, its OK if the inlet has a bit of flow separation at static conditions, as long as it cleans up below half the stall speed. Yeah, takeoff run is a bit longer, but not very much. I think the BVM sport jets do this. It keeps the inlets from looking too big. From my experience, the BVM DF F-86 does not fully clean up at full throttle until its above stall speed, but its a really tough design problem.

Note that for some of the VTOL fans I have been working on, at hover HALF the total thrust comes from suction on the inlet lip. If you have a sharp lip, not only does the rotor thrust go down due to feeding it garbage air, but the inlet lip thrust goes away entirely. This is a very very big deal! As speed increases, the lip suction goes down, and eventually the goal is to minimize spillage drag at high speed and lower throttle settings, but a nice rounded inlet can do that also.

For the VTOL airplanes, during hover the flow actually is running FORWARD along the sides of the body and into the inlet, for more than half the body length! As flight speed increases, the stagnation point (i.e. the point on the body that divides the air going into the inlet from the air going aft along the outside) moves forward, and then at high speed actually moves inside the inlet.

For the exhaust, basically, the exit area controls the engine loading. Smaller exit will slow down the engine. In general, more area is better (more mass flow means more thrust), but once you get into real hardware on ICDF, the critical thing is to get the engine properly peaked on the pipe.

I am not sure about the best shape for the tailpipe. There is a significant reduction in skin friction if you hold constant area for most of the duct then taper it at the end. However, for a rapid taper, I do not know how critical it is to get the duct walls parallel (or nearly so) at the exit diameter, or if you can just have a cone shape and chop it off.

"Area Ruling" is a bit of a misnomer. As mentioned above, its critical to keep the flow attached through the duct, so you want to avoid BAD changes in cross section area that cause separated flow. For the Viojett, bulging the duct to go around the engine is a good idea, but its also part of a very low drag, well thought out engine installation. However, for area ruling around a tuned pipe, where the cross section area changes are really gradual, its probably not that critical.

The BIG thing about duct performance is to minimize drag. Several things. First, the air goes through the duct quite a bit faster than the air goes around the outside. Drag of a given object increases as the square of the velocity. Even at top speed on a DF, the air is going quite a bit faster inside, so having some junk in the duct is going to have 50% more drag than it would if it was outside. It amazes me at all the people who make a DF that is super sleek outside, and then hang a lot of junk in the duct flow.

Second, junk in the duct not only causes drag, but that drag cuts the mass flow, which also cuts the thrust. A good rule of thumb is that the thrust loss is DOUBLE the drag of that object in free flow. So, if you add that to the effects of the higher flow speed, the performance loss of something in the duct is AT LEAST TRIPLE what it would be if you hung it on the outside of the model.

Third, for junk in the duct, things like linkages and fuel tubes are a big deal. Some of you may have seen things that compare the drag of a cylinder vs a good low drag airfoil. The wire or tube has a drag coefficient of over 1.0. A good airfoil can be under .01 (based on top view area). If you convert that to frontal area, the airfoil could be 15 times thicker than the cylinder for the same drag. So that 1/16" music wire pushrod going across the duct to the carb could have the same drag as an airfoil an inch thick! (note having a lot of sweep on the pushrod, like the viojett has, makes a big difference as compared to a torque rod going straight across).

Our models work in a very different aerodynamic environment than full scale stuff. All full scale jets have to deal with some amount of transonic or supersonic flow, and that changes all the rules. For example a key part of the design on the Harrier inlets was to handle the transonic flow around the outer lip at high speed. That meant a major compromise between good hover performance and high speed. Our models do not have that problem.

Same thing on interior flow. Full scale turbofans want about Mach .4 flow at the fan face, well below cruise speed. That gives a very different duct shape than we would want. Full scale airliner inlets actually design to get some supersonic flow at the inlet throat under takeoff conditions.. its a neat trick because the fan noise absolutely cannot propagate out the front! Again, we dont have that problem.

I guess the summary of that is do NOT blindly copy full scale ducting!

Finally, all this stuff is pretty critical for DF. They need a large air flow, and have a very very low pressure ratio, so even small pressure losses (i.e. drag) has a big effect. For our turbines, they have a lot lower mass flow, and a lot higher pressure rise, AND they have compressors that will tolerate a lot of turbulence in the flow, so you can get a way with a lot more. You also have a lot more performance to start with. However, some care in duct design will still pay off. (the biggest deal is probably with bifurcated tailpipes).

BTW, as an interesting data point, on one VTOL UAV, it is getting about 24 lbs of static thrust out of 3.5 HP from an 11.5" diameter fan. With that efficiency, a typical OS .91 and Ramtec would have about 13 lbs of static thrust. I don't remember the exact numbers, but my CyberHawk sport jet was measured close to that, so it is possible.

Bob Parks