[helloworld922]

I'm trying to simulate the pistons in an internal combustion engine. To start out, I created a simple tube for the chamber and added a cap on one end. At the other end I have a moving cylinder for the piston. I attached a simple reciprocating motion to the piston and assigned a variable air material to the empty space. The piston and chamber both have a solid aluminum material. When I run the simulation, though, I'm getting very wrong results. For starters, the pressure inside the chamber is way too small.

The piston is designed with a bore of 67mm, a stroke of 42.5mm, and a compression ratio of 12.8.

As the piston moves from BDC to TDC the static pressure changes from 0 kPa gage to ~60 kPa gage. I'm expecting the pressure to go up somewhere in the range of ~1200 kPa (give or take a few hundred kPa).

I've tried a variety of different methods to remedy this problem but nothing I've tried worked. Here's what I've tried (separately and in various combinations):

1. Increase the mesh density of the air volume
2. Increase the mesh density of the piston
3. Turn on heat transfer
4. Played around with the total temp. field (using compressible flow, all others crash the simulation). I tried values ranging from ~30 deg. C up to 1000 deg. C.
5. Played with different Inner Iterations values (1, 2, 10, 20)
6. Added various Boundary conditions to try and contain the flow. For example I put a 0 g/s mass flow as the boundary condition to the top face of the cylinder.
7. Added another solid aluminum cap beneath the piston
8. Moving the piston slower/faster (different ranges from ~60rpm up to 10000 rpm)
9. Using a fixed air material

Nothing seems to have really worked. Is there something I'm missing here?

I've attached a screenshot of what the piston looks like at TDC. The view is of a central static pressure cross section of the piston. The upper black section is the top cap and the lower black section is the piston. The surrounding 0 kPa section is the piston chamber (with additional bottom cap).

[jwdenhartog]

While the geometry may be simple, the physics in this one are complex. Depending on what you are ultimately trying to simulate, this model is probably best discussed with the CFD support team. They will be able to help you with the details of meshing and time step sizes (both are probably important here) and advise you of assumptions that will be made when the model is run.

http://subscription.autodesk.com/sp/...74&id=11607975