The actual paddle-pump idea has been around since forever. Usually primary and secondary rotors are different sizes.
The new part is getting it to actually work well as an engine. Modern ceramic rotors probably help.
The centre "hot side" bearing and oil sealing could be issues. - maybe ideas lifted from jet engines?
There have been jet turbine engines that used rotary recuperators before - that has a similar problem. (Rover).
<edit>
Is there a more favourable mechanism that would work in a similar way?
A downside is that inlet suction, compression are on one side, and expansion and exhaust are on the other side.
1) The exhaust side gets very hot, and will need some cooling, otherwise it is likely to want to run at 7-800 degrees c as an average between combustion temperature and exhaust temperature.
This is worse than with a turbocharger turbine.
2) The high pressure transfer process is likely to be quite lossy.
3) The loading on the rotary valve is quite high, and draggy.
4) The porting in the rotary valve is likely to be complex, and rotary valves are difficult to cool.
One part runs exhaust only, one part does intake, and one part does transfer.
5) There are likely to be inner walls between each side chamber and the rotary valve. These need cooling channels on the hot side as well.
The intake side likely doesn't need extra cooling.
6) Is there a full set of working drawings available? The devil is in the details.
7) this design could be adapted to 4 cycles on the one rotor. It probably doesn't need two sets of rotors, or a hot side.
It may need 3 sets of disc valves, though, and a side combustion chamber.
This would extend past the area of tooth engagement.
8) One disc could go the other side of the side chamber, and act as the exhaust.
9) One disc is needed to stop backflow through the combustion chamber;
10) One disc is needed for inlet.
11) Poppet valves are really better than combustion-loaded disc valves, with the sliding seal action.
This creates more drag than the cam, cam follower, and poppet valve system.
12) Has anyone tried to replicate the drum system that was evaluated by Mercedes racing? - it looks really complex.
13) I have urls of a combined drum/disc system, which I will post here.
This valving system looks relatively buildable - again, working drawings are not supplied, so you have to work out a lot of detail.
I suspect the disc/drums float, and actual sealing is on the disc side.
He claims that the bearings do not take combustion load, so The disc must be loaded to oppose exhaust pressure.
There doesn't seem to be any floating seals.
Side-on piston rings would make good seals, but the opening shape is not so good.
Lots of good stuff in this lot!
http://www.pattakon.com/pattakonPatRoVa.htmhttp://www.pattakon.com/PatRoVa/PatRoVa_I4_90_P1_cover.gifhttps://www.f1technical.net/forum/viewtopic.php?t=10966&start=1095
<edit>
Here is how a paddlewheel engine would work with 4 valves and a half engine speed camshaft, poppet valves,
and an offset combustion chamber.
Rotary valves would be tricky to include.
1) Requirements. - rotor clearances together and in the housing need to be very tight.
2) You need several gears, including a synchroniser set.
3) Could this be made to sufficiently tight tolerances?
4) is there a better rotor design that is more friendly to seals? - ie:
a) Sliding vanes - as in a vane engine. With an offset chamber, you need one moving vane and one fixed seal.
b) any other??
-Paddle tooth and a stop-start indexing wheel with seals. - Geneva mechanism? - image-
The indexing device can be arranged to move a pocket over the paddle once every cycle.
These devices are getting away from pure rotary action, though, and are likely to cause vibration.
The advantage is that you are not relying on two cylinders touching each at a point other to seal.
These would likely need spring-loading against each other to account for bearing tolerances.
5) How do you deal with sealing the sides and corners of the vane? do you use separate tip and side seals,
as well as round rotor seals?
6) You have the problem of a very long expansion "work" chamber.
Combustion should be complete before expansion starts.
7) Vane-type engines have not been historically successful.
Orbital Engines Co. did some work on a vane-type chamber.
