Stirlingengineforum.com

Jack wrote: ↑Fri Nov 24, 2023 11:38 pm If we're working with expanding and contracting air, and I want a closed cylinder, I can't use the one like in the picture you mentioned above. Because it goes back to nothing or very little airspace at some point in the cycle.
The displacer in my rotary design is just a minimum of volume of air being moved around. The lobe created by the off centre rotor is the working area or "piston" if you will.

So to compare it with the picture above, I would need to move the rotor further back to the centre. And the minimum circle it makes then is my displacer. Or all the working fluid I have to work with.

If I understand you correctly, since no air is actually entering or leaving through ports, you need to allow some extra room for the air to exist when at its maximum compression. That IMO relates to the compression ratio, or so it seems to me. Maximum expanded volume to minimum compressed volume of the working fluid in each chamber. Calling that air space allowance a "displacer" seems like a misapplication of the term.

Essentially there is no actual displacer.

It seems you could increase or decrease the compression ratio by the degree of offset.

Alternatively, maybe reduce the diameter of the rotor. Take it out and turn it down on a lathe?

Yeah maybe displacer isn't the right word for it, but essentially the space performs the same function. It moves my working fluid from hot to cold.

And you're right, those are ways to change the ratio. But it is still limited by the amount of heat it can absorb in the time it passes the hot area. So there is a limit to how much you can increase the air in there.

That's why I now switched to another line of thinking. Trying to pool the working fluid in a hot and cold area to give it ample time to transfer heat.

Jack, if I’m figuring your idea right, it might be similar to the working principle of the Ericsson engine. In that design there isn’t a dead-volume penalty for as large a heater as you want. It ran open-cycle so it didn’t need a cooler. It reciprocated and had a regenerator, but if you did run open cycle you could recuperate instead of regenerate by routing the exhaust air into the burner or perhaps to a water heater. That likely wouldn’t be as effective, but it’d be something. I like how having a cool rotor and a hot rotor eliminates thermal shorting problems. Anyway, if you haven’t already, you might look up the Ericsson cycle and maybe more likely the Brayton cycle, for some possible inspiration. I agree with Matt about rotary engine seals, but with a high enough speed and a low enough pressure difference the loss might be insignificant — maybe. I’d worry more about that low pressure difference overcoming friction, such that maybe looser/leakier vanes might be better — maybe. As to friction, your two rotors could be on one shaft instead of geared together. If it works at all, it should be capable of high rpm, so the generator could be on the same shaft too. I’m dubious about beating demon friction, but I’d like to see something that simple work.

Bumpkin

Thanks Bumpkin, it does seem like my idea evolved into a Brayton cycle. At least that kinda proves the physics of it, now getting it to work haha.

Jack wrote: ↑Sat Nov 25, 2023 12:37 am Yeah maybe displacer isn't the right word for it, but essentially the space performs the same function. It moves my working fluid from hot to cold.

And you're right, those are ways to change the ratio. But it is still limited by the amount of heat it can absorb in the time it passes the hot area. So there is a limit to how much you can increase the air in there.

That's why I now switched to another line of thinking. Trying to pool the working fluid in a hot and cold area to give it ample time to transfer heat.

Thanks for the clarification.

I don't mean to be disagreeable, but many standard Stirling engines run upward of 1000 RPM with much less surface area than this relatively large pie shaped rotary type design. What's your basis for determining there would not be sufficient heat transfer?

Also Bumpkin mentioned two rotors? One hot and one cold?

I saw something of that sort in one of the papers you referenced, but didn't think it was part of your design.

Anyway, I've always had an interest in rotary type Stirling as something that could have potential. You've given me some things to think about.

After some thought, I imagine the problem with a typical rotary displacer is not a lack of surface area for heat exchange but the way the volume of air interacts with the cylinder, or rather, the lack of interaction.

In most conventional Stirling engines the working fluid is driven forcibly into the hot surface perpendicular to that surface. This forces the air to travel in a direct collision course with the hot surface creating a lot of turbulence and interaction between the air and the hot surface.

With a rotary displacer on the other hand, the volume of air is mostly static, just sliding past the surface without much interaction at all, kind of like a box car on a train going through a tunnel, there is virtually no interaction at all between the air inside the box car and the walls of the tunnel, even if the doors to the boxcar are wide open. The surface area is there but the air passing across the surface parallel with the surface is not actually interacting with the surface but just sliding past going round and round with no turbulence to put the air into contact with that surface.

I've been thinking along the lines of adding some kind of fins or vanes to direct the air or create more turbulence but this approach doesn't seem to work. It just creates smaller static pockets of air that tend to just revolve around together, though I'm still thinking on it, I haven't come up with a solution.

That makes sense. I think in that respect the single rotor would be very limited.

And since realising I'm turning it into a brayton cycle engine I'm rethinking it all again.
Stirling engines are Delta t limited. Brayton engines are compression limited. So tit for tat I guess.
I quickly viewed a NASA paper that looks into combining the two, but that seems too complicated for my use.

So I think I'll stick to the dual rotor type in the paper I posted before and try to make one of those.

On the other hand, a rotary vane is not like a typical static rotary displacer. The air volume between the vanes is constantly changing and the working fluid is forcibly squeezed against the hot surface as it is moved along at least hitting the surface at an angle.

Also as the working fluid is expanded, that should set in motion some turbulence. The working fluid is not a fixed volume like the slug of air simply revolving round and round, so perhaps a rotary vane type design would be a solution.

A little more time and you guys should arrive here...



This georotor flavor has been kicking around for years. If you ever consider using vane rotors, best to play with a couple used Hydrovane compressor units, since these buggers can take a hefty cross vane pressure without failure (~150psi max). The vane pump scheme (and scroll pump scheme) has been well trodden in past and usually ends with Brayton Cycle failures. I get the idea...(1) Brayton uses adiabatic compression in one rotor and adiabatic expansion in the other rotor, so cycle is fast and heating & cooling are NOT supplied thru rotor cases (2) Brayton uses isobaric heating between rotors, so as Bumpkin points out, there's plenty of time to heat gas since heater 'volume' is largely independent of everything else (3) system can be open or closed cycle. Unfortunately, the obscure downside is that the intake mass to expansion rotor must equal the outlet mass from compression rotor for scheme to succeed. Look up the 1919 Stoddard patent for piston version of same scheme.

I almost forgot, this is last version where regen was added to impress treehuggers...

I'll mount a working piston on each side ...

Jack wrote: ↑Sun Nov 12, 2023 9:31 pm ...
I've been looking for a way to make the working piston rotary as well. It finally clicked.
...

I hadn't noticed (or forgot) about this earlier plan to mount the power piston(s) on the side of the displacer, which was causing me some confusion.

Anyway, I wanted to clarify that I had been thinking in terms of the later design since I started posting. Or....

Are you considering making the second rotor the power rotor while the other is a displacer rotor?

To further clarify, when I posted this:
I was thinking in terms of one rotor.

That includes your image with the vanes on the circumference:




Now I'm not sure if you considered this possibility. That is, the single rotor serving both or all functions.

My illustration does not represent anything like a displacer on one side and power rotor on the other but rather two single rotor engines sharing the same set of vanes, which was a spinoff from the single rotor design which is what I had thought you were talking about.

In that case the vanes themselves would act as "pistons" of sorts, in the same way they do in a rotory vane air motor. (That is, there would be no "attached" power piston or power rotor as the rotory vane itself would act as it's own power rotor).

This would work similarly to a rotory vane air motor but instead of driving the rotors by expanding the air volume with more injected compressed air, the air would be expanded by the addition of heat radiating from the rotor housing.

I think either design, using a single rotory vane motor, (as I thought you were proposing), or my double-acting rotor, might be worth considering.

At this point I'm thinking I may have gone off into a whole different ballpark without realizing it, having misinterpreted what you said: "I've been looking for a way to make the working piston rotary as well..."

I lept to the conclusion that you intended to use the rotor in the same way as a rotory vane air motor. i.e. as if you said: I've been looking for a way to make THE rotor serve as a working piston as well.

Now I'm really not sure what way you meant that but I can at least clarify what I had in mind.

matt brown wrote: ↑Sun Nov 26, 2023 1:16 am A little more time and you guys should arrive here...


Star Rotor.png

That, I think, is an internal combustion engine.

I'm wondering if the same design/concept could work as an external combustion engine?

The left side appears to simply be for fuel delivery, a kind of carburator/fuel/air intake, then an igniter so all that could be dispensed with leaving the right side:


But the intake and outlet ports could go as well as they serve no purpose in an external combustion engine.

The housing would then be a perfect circle with this offset "star" mechanism.

Heat the left side to expand the gas and force the rotation of the rotor, and cooling takes place on the right.

The question is, would something so simple actually work?

Personally I can't see any reason why not.

If the Star rotor scheme is Ericsson:

(1) one rotor case would be cooled compressor and other rotor case would be heated expander (2) transfer tubes would be crossflow heat exchanger (twin tube uniflow) due to continuous flow vs typical 'staged' regenerator (bidirectional).

However, if Star rotor scheme is Brayton:

(1) both rotor cases are isolated/insulated (2) heating is supplied in a first transfer tube between cases and cooling supplied in a second transfer tube between cases (3) some regen is possible, but limits scheme and introduces various issues.

You guys are missing one massive subtle difference between these two...with Ericsson, the gas mass exiting compressor will undergo isobaric heating AND expansion during regen between cases, but will largely remain a constant mass between cases due to heat balance between short 'blows'. But with Brayton, isobaric heating between cases (especially a large heater) insuring gas mass balance becomes a challenge. A simple diatomic (air) example is

(1) 6:1 adiabatic compression from 300k >>> 600k (1 atm >>> 12 atm)
(2) 1:2 isobaric heating from 600k >>> 1200k (12 atm >>> 12 atm where volume doubles)
(3) 1:6 adiabatic expansion from 1200k >>> 600k (12 atm >>> 1 atm)
(4) 2:1 isobaric cooling from 600k >>> 300k (1 atm >>> 1 atm where volume halves)

The problem with Brayton is keeping the gas mass constant between the cases (aka synchronized) while the 'volume' is changing during isobaric heating. The classic piston Ericsson nixes this issue since the regen volume is small vs the working cylinder volume, but an Ericsson with counterflow heat exchanger is leaning towards this Brayton issue. This mass control issue is was limits Brayton and why CCGT are a rare breed.

There is, BTW an animation for this engine on the website:


https://www.starrotor.com/engines/


Which is kind of interesting.

An apparent hitch is that the rotor migrates around from the hot side, eventually making its way over to the cold, which might make cooling and contraction of the working fluid after power output difficult if not impossible.

My experiments, however seem to indicate that the working fluid cools and contracts due to power conversion, (heat converted to work) so that it will cool and contract after power output regardless of the housing or power cylinder temperature.


https://youtu.be/LG09AXAjpio?si=lOLVXAsp2UVGOx1A


If observations of this kind are really true then this would open up new possibilities for engine designs of this sort.

Theoretically, in my mind at least, a rotor that maintained some heat would increase power and reduce loses.

Tom Booth wrote: ↑Sun Nov 26, 2023 9:48 am Now I'm really not sure what way you meant that but I can at least clarify what I had in mind.

I've been evolving my idea pretty fast the last few weeks and posting as I go along. At first I thought of a separate rotary displacer and piston, but then I landed on the same idea you're proposing. A single rotor that's doing it all.
So I understand what you're talking about.

Right now though, I'm leaning more towards a dual rotor system on the same axle.
One main reason is that heating one side of a cylinder will warp it and that will affect the sealing by the vanes.
Another is separating a cold and hot rotor increases the delta t.

It allows for a little more design freedom to increase efficiency.