CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FloEFD, FloWorks & FloTHERM (
-   -   Simulation of a regenerator in stirling engine (

Walle March 8, 2013 07:05

Simulation of a regenerator in stirling engine
Hi everyone! 'S been a while :)

I want to simulate a regenerator for use with a stirling engine (alpha type), but I don't know if it is possible to simulate it the way i want. I'm using Solidworks FlowSimulation package (not the express).

The cycle i want to simulate is like this (for those of you unfamiliar with the stirling cycle):

1: Hot gas with an elevated pressure flows from the hot cylinder through the regenerator to the cold side.
2: The gas then drops in pressure and flows back through the regenerator to the hot side.
(I know this is simplified a lot)

I don't have to (or even want to, in this stage) simulate the actual cylinders heating or cooling of the gas, to start with i could even omit the pressure changes. But is there a way in a time dependent simulation to reverse the flow of the gas, and at the same time also change the temperature of the gas? The only way I can think of right away would be to have two boundary conditions on each side of the regenerator, and make those time dependent. But that will give unwanted seccondary effects, plus I'm not sure if I can give it a sinusoidal motion (the gas doesn't just switch direction, it's pushed by pistons, so the flow of the gas should be pretty sinusoidal).
The regenerator I want to simulate would be on the tube connecting the cylinders to each other.

Any ideas on how I could set this simulation up? The goal here is to pretty much compare efficiencies with different types of regenerator design.

Also, I would like to know if it is at all possible to simulate the whole system? I know there is some kind of option for moving walls, but could that be used as a piston in this case? Or is it possible to do a simulation where the piston "sucks in gas" and then pushes it out again, to see how much the gas was heated (or cooled)?

Edit: I can answer some of my questions:

I can probably simulate the regenerator with only one set of BC's (i.e. one BC at each end of the regenerator) and using transient analysis with table or expression driven pressures for the BC's. The tricky part will be finding the correct pressures to simulate with, but that problem lies outside of the scope of this CFD-problem.

The moving wall cannot be used to simulate a piston, since the actual mesh cannot move. Instead, the moving wall option is to simulate friction effect of air-surface. One example would be simulating the drag of a car. Then one should specify the "road" as a moving wall, otherwise the simulation will rather be wind blowing on a stationary car.

Boris_M March 20, 2013 05:49

Hi Walle,

this is an interesting project you have there.
Yes, you can use pressure and try it this way to change the pressure either by a equation or table. But as you said you will need to know the pressures.
Another way which is a little bit easier as you don't need to know th pressure but will need to create two steps in two projects. You would use the one flow direction in the one project and use the final results of the first project as initial conditions in the second project where you change the BC to reverse the flow. Such a change in the initial conditions can be done in the general settings. there is a drop down in the initial conditions where you can switch the parameter definition from "user defined" to "transfered".

For the piston, this is true, there is no moving geometry but you can try to do the volume flow also transient as you can describe it by the cylinder speed and dimensions.

I hope this helps,

liuzexiang September 14, 2013 09:32

i have no idea, but i think it is very i will pay attation for this.

All times are GMT -4. The time now is 13:23.