Hello All,
In molflow,I can set “part diffuse” and “part specular”.For example,I set 0.8 part diffuse and 0.2 part specular. In my understanding,the reboud particles’ direction will have 80% follow cosine law and 20% particles after the collision, the momentum exchange occurs just along the normal direction and the reflected gas lies along the specular direction. And for 80% particles,their velocity will completely follow Maxwell distribution,if I don’t set accomodation coefficient(default Acc is 1 as mentioned in molflow doc.).
But if I set part diffiuse and accommodation coefficient simultaneously,does that mean diffusion part particles’ velocity will not completely follow Maxwell distribution?It will follow the equation provided by molflow doc below.
If so,there has some different between rebound model which molfow use and Maxwell-reflection model.In Maxwell-reflection model,there also has a thermal accommodation coefficient Sigma,But this Sigma seems same as the value set in “part diffuse”.I will paste the article below.
It seems like molflow’s “part diffuse” and “part specular” are Maxwell-reflection model.And the thermal accommodation coefficient is different with Maxwell model? The thermal accomodation coefficient in molflow is an adjustment of rebound speed?
Thanks in advanced.
The “Accomodation coefficient” only affects the particle speed, and not the direction:
The v_wall follows the Maxwell-Boltzmann speed distribution characteristic of the wall’s temperature.
The diffuse/specular/cosine^n settings, independent from the velocity, control the direction.
Specular is perfect mirror-like reflection
Diffuse is random, cosine-like (also called Lambertian) reflection
Please note that in the general case, the momentum exchange also happens through the parallel, not just the perpendicular component. If you would shoot a gas jet near parallel to the surface, which would reflect it diffusely, you would see a molecular force acting on the wall in the jet’s direction.
Hello Marton,
As your reply,the fractions of diffuse/specular/cosine^n settings are different from Accomodation coefficient in molflow. The fractions affect direction,Accomodation coefficient affect speed.
But in molfow’s doc references, there seems only one value.And in other papers,I also see this conslusion:
Alpha part of molecules are diffusion and with Maxwell speed.That is, the thermal accomodation coefficient only determines the direction of the particles after collision, and the velocity still follows exactly the Maxwell velocity distribution, which is not the same as molflow.
In molflow, the coefficients that regulate how much of the molecule is specular and diffuse are two different values from the thermal accomodation coefficient. Molflow has an additional “thermal accomodation coefficient” that regulates the velocity of the molecules after collision than in the literature. But the value of the “thermal accomodation coefficient” is the same as the thermal accomodation coefficient that regulates the direction of the molecules after collision in the article, which partially confuses me.
Although I can solve this problem by inputting the thermal accomodation coefficient from the article into the diffuse/specular/cosine^n settings section of molflow and keeping the “thermal accomodation coefficient” in molflow at 1 to make molflow consistent with the article, I would like to know the reason for molflow to add an additional “thermal accomodation coefficient” to regulate velocity. I would like to know the reason why molflow has a separate “thermal accomodation coefficient”.
Thank you! It looks like I cited this article but didn’t spot this controversy (at the time of writing my thesis, I believe inly fully specular or fully diffuse reflection was possible).
Cheers, Marton
Hello Marton,sorry for late reply.I left school for few days.
Now I meet a new problem.In moflow I could set "part diffuse"value dependent with time.However,is there any solution to set “part diffuse” value dependent with pressure?
Thanks!
I don’t think “part diffuse” can be set to be time-dependent.
Molflow can’t have any value depend on a simulation result (i.e. the pressure), as it’s not simulating in chronological order.
According to paper I read, accommodation coefficient dependent with pressure,so I want set a P-dependent parameter to simulate how P effect molecule—wall interaction.
