Body

I would like to simulate "quality" of the molecular beam effusing from a channel. "Quality" of the beam in my case is not only its incident angle and orthogonal velocity distribution, but also tangential velocity distribution. I would be happy if you added this "tangential velocity" option to the profile of a facet.   

Submitted by Marton Ady 6 years ago

In case of the orthogonal velocity, the direction is straightforward. In case of tangents, the azimuth angle can be anywhere on a 2PI range. Just to be clear: you'd like a tangential velocity magnitude distribution?

More concretely: if two molecules hit a facet near-parallel to its surface, in the opposite direction, should they cancel out (vectorial addition) or count as the same (scalar addition)?

If you're looking for the latter, can't you derive the tangential distribution from the incident angle distribution and the average molecule speed on the facet? (It is unlikely that the molecule speed would be angle-dependent)

Submitted by 0b5f164b4a0795965d0c 6 years ago

Indeed, I would like to have tangential velocity magnitude distribution using scalar addition. Unfortunatelly I don't know how to get it out of the incident angle distribution and the average molecule speed distribution. Could you give a hint? Or if I could get the raw data of the velocity vector of each incident particle on the choosen facet it would be obvious for me how to calculate tangential velocity magnitude distribution.

Submitted by Chao LI 5 years ago

 

Hi Marton,

The tangential velocity distribution is for magnitude distribution, right? You did not do vectorial addition, right?

The tangential speed distribution distibution is likley angle-dependent if you have multiple point-like emitting sources hitting a certain dectecting region. That would be wonderful if we can specify the step of azimuthal angle for the tangential speed distribution (the tangential velocity magnitude distribution), like what you have for the incident angle profile.

Submitted by Chao LI 5 years ago

It seems that the user manual is not up to date:

https://molflow.web.cern.ch/sites/default/files/molflow_user_guide.pdf

 Any instructions about the setting for incident angle distribution? How do you define Phi? What is the origin of Phi and where does it end? I mean the coordinate system for your Phi.

Thanks. 

Submitted by Chao LI 5 years ago

I guess I've got the answer for the coordinate system. Once I click to show the u and v vectors with normal vector, then the azimutual angle and polar angle are defined as convention.

Submitted by Marton Ady 5 years ago

Answers to your questions:

1) Tangential velocity distribution is magnitude-only, opposite vectors don't cancel out.

2) All three profiles (incident, orthogonal and tangential) are phi-independent.

  • Incident calculates theta elevation (angle between facet normal and incident direction), 
  • Orthogonal calculates the velocity component towards the facet normal
  • Tangent calculates the magnitude of the component perpendicular to the normal, regardless of orientation (changing U,V should not change the tangential distribution)
Submitted by Marton Ady 6 years ago

Ok, I've made a temporary version for you, download it here:

molflow_2.6.63_tangential.zip

When you make an orthogonal profile, it will in reality count the tangential component.

Incorporating it properly (interface and file format changes) will take some time so it will be included in the next version.

Please note that there is a difference between the distribution of hits on a facet and the distribution of (a volume of) the gas, therefore the "Surface->Volume conversion" checkbox in the profile plotter. I detail it in section 2.8.0.5 and 2.8.0.6 of the algorithm documentation. In case of the tangential velocities, such normalization doesn't work (the tangential component doesn't affect the hit frequency), therefore keep it turned off.