Tverskoy-Pasechnik Rotary Steam Engine

[russkey]

This is the rotary steam engine I am currently designing. It is based on the Tverskoy steam engine patent from the late 1800's. Not too much to go on in the available documentation, so everything beyond the basic sketch is my own work. I have attached an image of the main four components (drum with paddles, two rollers, and a cylinder casing), as well as an animated gif of the components in action. I'm currently working on the steam distribution system, which is all contained within the various channels, holes, and inserts of the drum and cylinder covers. The paddles on the drum act as the piston surfaces. The rollers keep the steam from running to the opposite half's exhaust port. The cutouts in the rollers allow the paddles to pass them in the dead zone. I am still looking for a way to have a tighter fit between the rollers and the paddles. Since the paddles are not in a pressed state at any point, they should not wear down like those found in other rotary designs. There are some Russian power transmissions based on this principle (drum+paddle+rollers), as well as flow meters. The transmissions use water or oil as the working fluid, instead of steam, meaning that this design can also be used as a pump. Comments?

[SL Ethel]

Looks like a neat design. You will definitely have the only steamboat with a Tverskoy-Pasechnick.

I would love to know more about several aspects of your design:

- Where does the exhaust port connect? Do you try to get uni-directional steam flow and associated efficiency advantages similar to a uni-flow engine?

- How are you doing seals? This has always been the sticking point (literally sometimes) for non-turbine rotary engines. It appears that you will have a lot of rubbing surfaces that will need very tight, yet low friction seals. What material will you use?

- And related to the previous question - what pressure and superheat do you plan to run at?

- Are you going to build a model size first, or go right to the full size engine. And what is full size, in your case?

Cheers,
Scott

[russkey]

- Where does the exhaust port connect? Do you try to get uni-directional steam flow and associated efficiency advantages similar to a uni-flow engine?

Looking at those diagrams, take the horizontal centerline. Just above the centerline near the left roller and below the centerline near the right roller are the input ports. Just above the centerline near the right roller and below the centerline near the left roller are the output ports. So this is basically a uniflow engine, right? There are effectively two "cylinders", one is the top half, the other is the bottom half. Steam enters on one side, and travels clockwise to the other side.

- How are you doing seals? This has always been the sticking point (literally sometimes) for non-turbine rotary engines. It appears that you will have a lot of rubbing surfaces that will need very tight, yet low friction seals...

This is a tricky point indeed. As to the perimetrical surfaces, my guess is that if I machine them 20 microns or so oversize, they'll wear down to just right after a couple of hours of operation. Once they are out of contact, they should not wear anymore, since there's no particular reason for them to touch. That said, this will partly depend on the tolerances of the bearings I use for the three axles. As for the upper and lower surfaces (the faces facing towards and away from the views in the images), I'm not sure. I hope that I will be able to devise some sort of load bearing which will keep them close, but not touching. Its one of the design problems I'm working on now.

- Are you going to build a model size first, or go right to the full size engine. And what is full size, in your case?

The images you see are the SolidWorks renders of a model engine I'm working on. I plan to machine it out of aluminum. The rollers are 1.5" diameters and the drum without the paddle is 3" diameter. (Note the 2:1 ratio, combined with the 2:1 rotation ratio of the rollers and drum means that there is no sliding friction between them, just rolling/touching friction until they wear down to size (timing is handled by external gears)). I picked aluminum because its easy to machine and fairly cheap (lots of scrap I can use at the lab and plays well with our CNC mill). A larger scale version would probably be made with brass (for the low friction) or steel and have a drum diameter of about 12". I'm also not sure about the pressures etc. yet. I'm planning on starting the model off on compressed air to see if it even turns. Before that I'll need to sort out port and channel sizes. Once I know it actually spins, I'll throw in some steam and see how it behaves.
I'll definitely be posting my progress along the way (and asking lots of questions).

[DetroiTug]

Interesting design in 2D.

Just an opinion...

In practice, sealing surfaces would be difficult to achieve and nearly impossible to maintain to result in any sort of stable performance. Even small clearances would add up to large areas of blow-by. These clearances would change with temperature. The interior profiles would have to be ground. Sealing the sides of the rotor against the housing would have to be resolved as well.

Thermal expansion of all the components, including the housing would be it's most considerable nemesis.

Reliance on timing between components would be very critical, with zero lag (deflection of components under acceleration) and backlash. Gears, timing belts, nor chains would be adequate. They all suffer deflection of components and backlash. Gears cannot run with Zero clearance at high RPM and fluctuating temperatures. (For long anyways)

Don't know the depth of the housing, but the face of the paddle will dictate the torque in relation to the pressure. If the paddles need to be lengthened, valve diameters increased it will exacerbate the issues.

Some things to consider..

-Ron

[russkey]

Ron,

These are all perfectly valid points that I will definitely have to consider in the design. What do you mean by interior profiles having to be ground? My plan was to CNC mill all of these parts. There have been functional prototype internal combustion engines based on this drum/paddle/roller principle, and they would presumably be dealing with much higher temperature gradients and pressures. This gives me hope that, though difficult, these problems are not insurmountable. So, fingers crossed and pencils sharpened!

Al

[87gn@tahoe]

You could use Apex seals for a Mazda "Wankel" rotary engine on the vanes.

[Mike Rometer]

Please allow me an observation! The problem with the Wankel rotary engine was indeed the rotor seals (both with the face and ends). I suspect the main reason for that is that on the rotor face there was basically only one sealing edge. In a piston engine each ring provides a seal. Could there be a method on providing more than one sealing edge for each lobe that is in contact with the "bore" at all times. I also believe that the Wankel relied largely on centrifugal force to keep the seals in full contact.

The major benefit of any new design must be that it is ether more efficient in manufacture or in use, preferably both. Of course in our case we can have fun just trying something defferent as a one off.

[DetroiTug]

Why would the internal profiles need to be ground?

Not sure what type of CNC machine is being contemplated for the manufacture, but double-digit micron tolerance on some of the best machines available would be nearly impossible to hold. 20 microns is equivalent to .0007". Holding that with Arc interpolation errors, as many CNC controls work on approximations in their internal calculations, that can create cumulative errors throughout an arc. Also factor in tool deflection, tool diameter, tool length, material hardness, tooling sharpness can all have an adverse effect on precision.

Another thing to consider from the drawings posted: The tip of the paddle must fit the radius of the main chamber, then it is going to have to somehow change geometry to mate and seal with the internal radius of the valve which is a different (smaller) radius. This transition is would have to rely on perfect timing between the valve and the rotor, approach and depart. Even clearances between the valves and housing.

Please don't take this input the wrong way, hopefully I'm giving you some other things to consider.

-Ron

[87gn@tahoe]

The earlier seals in the Wankel engines were of issue due to how narrow they were and their composition. The current seals available will withstand 1600*F as well as the combustion pressures of a engine singing at 10,000rpm and 30psi of boost.


I think in a steam engine even the original design seals would have little problem.

[russkey]

I presume that there was a time when even 50 micron machining accuracy was not possible, engines had to make do with some sort of tolerances and gaps between the cylinder and piston. Are there guidelines for an acceptable measure of seal imperfection? I would imagine that it might depend on the pressure difference between the parts being separated? Sorry if I'm jumping the gun here and should just wait for my book to arrive with the answers to my questions.