Simulating flow with Direct Simulation Monte Carlo

[Declan]

Hello everyone!

So far I am simulating a Maxwell-Boltzmann distributed gas at 273K in a cube. The walls are thermal but also set to 273K. So the temperature should stay the same throughout the box, which it does.

But now I'm trying to get the gas to flow. I'm trying something, but I'm not sure if it's the right way to go about it. First of all, I now distribute the velocities almost the same way, but I add an "offset speed" to all of them, so they have the same basic distribution, but shifted (ie, the center of mass is moving now, rather than stationary as before) so they'll be moving towards the "back wall". I have made the "back wall" kind of a one way valve: if a molecule hits it, it's gone forever. I have made the "front wall" kind of a pump: if molecules hit it, it acts like the other normal walls. But I add a particle every time step (to make up for the ones leaving) pretty close in front of that wall (but not directly on it, because that causes problems). I distribute the particles the same way I originally initialized the original particles, but I again add an offset speed, towards the "back wall" (the one way valve, the exit, etc).

But my measurements are being troublesome. The temperature is lower than it should be. The average x velocity should ideally be the offset speed. Near the place where particles are added, it's much lower. Near the back wall, it's much higher.

I know I'm probably not doing this the way most people using DSMC do it. I've googled a bit but couldn't find much. Could someone shed light on what I might be doing wrong? Or show me a paper that explains how to properly set up flow in DSMC?

Thanks!

 

[eljose79]

Hi there! It sounds like you have a good understanding of the Maxwell-Boltzmann distribution and how to simulate a gas at a constant temperature. However, your approach to creating flow in your simulation may not be the most accurate or efficient. In DSMC simulations, the gas particles are typically given random velocities based on the Maxwell-Boltzmann distribution, and then allowed to collide and exchange momentum with each other as they move through the simulated space. This approach more accurately models the random motion of gas particles and the flow of the gas.

In your current approach, you are artificially adding an "offset speed" to the particles, which may be causing some inaccuracies in your results. Additionally, having a one-way valve and adding particles at the front wall may also be affecting the flow in your simulation.

I recommend looking into papers or tutorials on DSMC simulations to see how others have set up flow in their simulations. You may also want to consider using a software specifically designed for DSMC simulations, such as DS2V or dsmcFoam.

I hope this helps and good luck with your simulation!