Sliding-Vane ICE

My inspiration for this idea comes from a design for an air compressor I've heard of.

In this compressor, there is a central hub with radial slots cut into it. Each slot is fitted with a spring-loaded vane. This central hub with vanes is located off-center within a larger chamber, which has a circular or elliplical cross-section. The hub rotates, and, as it does, the vanes slide against the walls. Because the hub is off-center, the distance from the center of the hub to the edge of the chamber varies as the hub rotates, so the vanes are pushed in and out over the course of a rotation. The volume between vanes (a "wedge" of air), therefore, increases over one half of a rotation and decreases over the other half. In the air compressors, there is an inlet port at the point of largest volume and an output port at the point of smallest volume, and that's all there is to it.

Half of a revolution, in other words, is intake/expansion, and the other half is exhaust/compression. So let's combine two of these compressor devices to perform a total of four "strokes": Air/fuel mixture enters the first device, which performs intake and compression functions, and then moves into the second, which performs combustion/expansion and exhaust functions.

Here's a more detailed description: Imagine two of these air-compressor-like devices stacked atop one another and attached to a common shaft. In the top half of the engine, the intake port extends over nearly an entire half of the rotation (the expansion half), and, as the vanes are pushed out against the walls and the volume of each "wedge" increases, air is sucked in. Once that half of the revolution is done, we get to the compression half. Naturally, the intake port ends at the beginning of the compression half. As the vanes are pushed in and the wedges are made smaller, the air is compressed.

Then, shortly before the point of greatest compression, there is a large port leading to the second half of the engine.

Now we're in the second half. As soon as rotation brings the current "wedge" of gas past the port from the first half, it comes to a spark plug, which ignites the fuel-air mixture. The gas then expands over this first half of the revolution. Then, beginning at the point of largest volume, and extending until shortly before the port from the first half of the engine, there is an exhaust port. During the compression half of the revolution, exhaust is forced out.

I'd expect this engine design to have many advantages over traditional piston designs. First, all strokes of the engine are always performed simultaneously, so combustion is always occuring, thereby eliminating the need for a heavy flywheel (though a small one may make sense to ensure smooth operation). Also, it is a relatively compact design with a large working volume, so it should have excellent power-to-weight characteristics. And with all these wedge-shaped chambers, I would expect pretty explosive torque characteristics.

I do have a few concerns though. First, there's the fact that these vanes are compressing and decompressing springs repeatedly and rapidly. At first I though that this would quickly cause fatigue. However, isn't this exactly what the valvesprings for the poppet valves in a traditional automobile are subjected to, without incident? So I think that concern may be a nonissue.

My second concern is sealing. The seals would need to be essentially the same as the apex seals used in Wankel-type designs, which have proved problematic in the past. However, those problems have been solved in Wankels, and this uses sealing which is nearly identical, so I suspect that this too is already a solved problem.

Now, this idea is also easily adapted to two-stroke operation using a single half of the 4-stroke version. Scavenging would be easy to achieve by overlapping the intake and exhaust ports slightly, and by placing the intake towards the center of the engine and the exhaust towards the edge; centrifical force would facilitate scavenging.

An external-combustion Stirling-like version would also be simple, again, using just a single rotor like the two-stroke version. One need only put the expansion half of the engine in contact with a hot reservoir and the compression half in contact with a cold reservoir. One could run such an engine using the temperature difference between an external burner and the atmosphere (using a heat sink), or between the atmosphere and liquid nitrogen.

In other words, the concept seems very versataile, and isn't even limited to four-stroke operation.

So now, my question is: What's the catch?

[Edits: Subsequent Changes to idea]

- Springs are unnecessary to press the vanes against the housing; since its spinning there is centrifical force.

- Rather than a two-part engine, one can achieve four strokes from a one-part engine consisting of a rotor centered in an ellipse!