Good points, Graham. Thank you. In fact, thank you all, guys. Let’s keep this discussion open. Even if this engine does not get built and does not wake up the entire Oxfordshire with a start sometime this year, to mark the start of the Fire and Noise World Conference, it may provide an entertaining and possibly useful exchange of opinion and knowledge.
Judging by the first reactions, people like the basic idea of SP-14, and I can now offer it with somewhat greater confidence as public property. I claim no rights to this concept other than the right to call the initial configuration published here ‘my idea’. Let the rest be the property of people in this forum, if they indeed decide to take it up. I will be very proud indeed if it happens.
Graham C. Williams wrote:Perhaps consider two things. 1) The cooling of the internal surfaces and struts. Past work featured on this site has shown that these may melt. Perhaps run the fuel through them or add some air holes.
I would normally try to use the internal annular cavity (the one that resembles the annular wing) for some purpose. Instinct tells me I should use it as a vaporization chamber for propane, which should – I guess -- be the fuel of choice during the prototype development. Vaporization and subsequent heating of propane should serve to cool the walls well enough. What I need is a trick that would let the vaporized (high-pressure) propane stream into the inlet annulus during the intake phase but not during the expansion phase. I may have a trick or two up my sleeve. We’ll see.
The internal brackets may be a tougher proposition. Given that they are likely to be small, perhaps they could be fabricated out of inconel. Perhaps a small sheet of inconel won’t be terribly expensive. Scrap sheet-inconel parts may also be available.
If that proves impractical, I intend to turn to my old trick of making one of the tubes fluted instead of cylindrical. Consider two concentric tubes. Look at their cross-section. If you make either the outer or the inner one longitudinally fluted, so that its cross-section resembles a petalled flower (or a Dynajet valve stamping) you can have one of the tubes cling to the other at the projecting points. The inner and the outer tube can thus be connected without discrete brackets.
By playing around with the exact design, you can achieve a preservation of the original section areas, too. Of course, you end up having a greater surface-to-volume ratio, but if one of the tubes carries cool air for augmentation purposes, it need not be a bad thing. The configuration may also help augmentation by offering superior mixing of hot gas with cool air. It will also lower the noise level.
Graham C. Williams wrote:2) At the point where the induction pipe becomes the combustion chamber the design might benefit from a greater discontinuity in the lines.
Graham, the diagram is not to scale. Also, circular sections easily deceive the eye. The farther out on the radius, the greater the deception. Remember that the annular intake passage curves outwards into the combustion zone from the position already far out from the engine centerline. That means that the small-looking inner radius of the transition from intake to the combustion zone actually represents a considerable rate of increase in cross-sectional area. If you drew an equivalent tubular Lockwood, you would get a very sharp transition indeed – certainly sharp enough for flow to detach from the wall.
This will be helped by the tendency of the flow to cling to the inner wall, whose slope is almost imperceptible and may actually be eliminated completely. I am counting on the inner wall ‘guiding’ most of the flow towards the curving ‘bottom’ of the combustion zone, which will force it outwards, acting as a centrifugal diffuser.
The inflow will (I hope) curve rather sharply into a tight annular vortex while its dynamic pressure will rise. This vortex will spin away from the intake and towards the exhaust. Wave reflections off the exhaust end will act to slow this rotation down, I guess, but I hope they will not stop it completely. Even if they do, however, the energy of rotation will translate into a pressure increase, so that any loss there should be minimal.
My worry is that the initial vortices – those formed in the intake tract as the intake air crosses the intake lip and enters the tract – will necessarily be in the opposite direction. I have no idea how to resolve this at the moment. Maybe the overall result will just be a lot of turbulence, rather than a distinct torus.
Graham C. Williams wrote:Dimensions. Only my first thought on dimensions this but consider it as a crossection of a 'Lockwood type' motor, displaced some distance from the centreline, then integrate through 2pi about that centreline. All the area and dia. ratios would have to be the same at each station - Would that work?
If I understand you properly, that is exactly what I plan to do. I will use the dimensional relations given by Kentfield for one of his Lockwood-style combustors – either the one on which he performed experiments with liquid fuel injection ort the one on which he tested the noise pattern. However, perhaps the Ecrevisse dimensions also deserve attention.
The starting point for all calculations will be the area of the intake tube mouth of the regular tubular Lockwood. From that, I can calculate the optimum area and diameter of the fitting augmenter. This will be the dimension for the inside tube (central duct) of SP-14.
It is also the outer (larger) wall of the annular intake tract. Deducting the intake section from the central duct section will give me the section (and then diameter) of the air intake into the central duct.
This diameter is also the inner diameter of the annular combustion zone etc. etc.
I am really interested in the behavior of the curious inner augmenter. Namely, I have never seen any studies of reverse ejectors – where the ingress of the driving fluid is annular while the driven fluid is in a cylindrical central flow.
Graham C. Williams wrote:Did you try this idea with the Sanders-Roe valveless? The induction would then be on the outside and the exhaust would blast into the central air passage. This central air passage could then be arranged as a diffuser; with a half angle of about 4 to 7 degrees this should increase the pressure and lower the cold gas velocity near the exhaust. Perhaps this might increase the forward velocity over which reverse flow takes place?
An interesting idea. I will consider it. However, when I turn the Saunders-Roe inside out, I do not get the same configuration as you do, judging by your next words about reverse flow: “Reverse flow from the induction could then be treated as if flowing over a plug nozzle or aerospike, a simple cone with a 30 degree half angle and extension tail.â€