another questions about Cylinder.m

Palabos General
#1

Hi Dear all

I read this program a lot. but I have some question about it:

  1. I read this document Unit Conversion for unit conversion. it is very useful. if I understand it, stage of LB unit conversion is:
  2. set u[sub]lb[/sub]~.02
  3. choose (Delta[sub]x[/sub]= 1/(No. of node) it is not physical value like .0001 meter , L[sub]0,p[/sub] , t[sub]0,p[/sub] )
  4. calculate Re No., Delta [sub]t[/sub]
  5. Run Simulation
  6. U[sub]p[/sub] = L[sub]0,p[/sub] / t[sub]0,p[/sub] * u[sub]lb[/sub]*Delta [sub]x[/sub]/Delta [sub]t[/sub]
  • is it true?
  • in cylinder.m we use : Umax.
    what is it? U[sub]lb[/sub] at entrance?
    in Unit Conversion i read that u[sub]lb[/sub]~.02 is good for u[sub]lb[/sub].
    but in cylinder.m we use 0.2. I think it is true because less than C[sub]s[/sub][sup]2[/sup].
    Thus that suggestion …?
  • Why it is not Sphere.m?

Best Regards

#2

Yes, it seems that you got it right. And, affirmative, the velocity in the Matlab example is a bit high (I think it’s 0.1, actually). This doesn’t mean that the simulation is wrong, it just mean that it isn’t exceptionally precise because the accuracy is impaired by compressibility errors. On the other hand, the point of view I like to adopt with Matlab is that we are not seeking high accuracy anyway. It is more of a prototyping tool with which parameters and models are explored before we turn to production mode with a more efficiency-oriented language. Given the performance of Matlab scripts, high resolution grids cannot be used, and it would be useless to waste computational time by turning down the Mach number beyond a certain value.

The rule of thumb “ulb~.02” is nothing more than that: a rule of thumb which provides you with something to start with. In practice, the right thing to do is to play with this parameter until you get a numerical accuracy corresponding to your needs.