Re: [ESPResSo-users] lbboundary

[Ulf Schiller]

Hi,

you need to make sure that h*c_s^2/\nu is small to avoid nonlinear
instabilities. h is the LB timestep, c_s is the speed of sound, and \nu
is the kinematic viscosity. In the D3Q19 model, c_s^2=1/3*a^2/h^2, so
a^2/(3*\nu*h) must be small. It may work with values O(1) but it is not
guaranteed.

In general, it is better to think in terms of dimensionless quantities
such as Reynolds number, Mach number, Schmidt number, or what have you.
These are the quantities you need to match in order to get the physics
you want.

Cheers,
Ulf

On 05/12/14 06:09, roya moghaddasi wrote:
> My MD timestep is 0.005, considering this can I use 0.025 for LB
> timestep with the aforementioned parameters?
> 
> Best,
> Roya
> 
> 
> On Friday, 5 December 2014, 9:11, roya moghaddasi <address@hidden> wrote:
> 
> 
> Dear Stefan,
> 
> Thank you for your time and consideration.
> I changed my parameters to this:
> lbfluid dens 0.362  visc 3.8  agrid 1.0  tau 0.05
> lbfluid friction 14
> 
> What is the relationship between viscosity and absolute timescale?
> Best,
> Roya
> 
> 
> On Thursday, 4 December 2014, 22:55, Stefan Kesselheim
> <address@hidden> wrote:
> 
> 
> Hi Roya,
> 
> On Dec 4, 2014, at 8:04 PM, roya moghaddasi <address@hidden
> <mailto:address@hidden>> wrote:
> 
>> Hi Espresso users,
>>
>> I want to use a cylinder boundary for my lbfluid. I want to ask some
> questions:
>> 1. can I use penetrable or direction or reflecting for this boundary?
>> 2. If I use "reflecting 2", will the boundary condition be bounce-back
> condition which is nice for a lbboundary? What is your suggestion?
> 
> Both options exist only for constraints and not for lbboundaries.
> 
> 
>> 3. Espresso manual says "This has important implications for the
> location of hydrodynamic boundaries which are generally considered to be
> halfway between two nodes to first order". my main cylinder which
> contains all the particles has a radius of 25 and I want my lbboundary
> to have the same radius. Do you see any contradictions of this and the
> Espresso manual statement?
> 
> 
> lbboundaries use the "link bounce back" method, that is described a lot
> in the LB literature. If you create an lbboundary cylinder with a
> certain radius around a certain point all nodes inside this cylinder
> will be flagged "wall". If the walls were plane, The the hydrodynamic
> boundary appears halfway between the outermost "fluid" and the innermost
> "wall" node.  The lbboundary node positions are half-integer, i.e. 0.5,
> 1.5, 2.5 …
> Thus, if you make e.g. a cylinder with radius 10 around position x=15,
> node x=4.5 will be "wall", and node x=5.5 will be  "fluid". The
> hydrodynamic boundary will be at 5, thus you get what you expect. You
> should, however, check the results i.e. by applying a pressure gradient
> against the analytical solution. From you other mail, I see you might be
> using a viscosity that is not suited for using the link bounce back
> method because it is too high.
> 
> Note that in most soft matter applications, the absolute value of the
> viscosity does not matter anyways, because the fluid is described by the
> linear Stokes equation and changing the viscosity only changes the
> absolute timescale, but that can be corrected for at the end.
> 
> Cheers and good luck
> Stefan
> 
> 
> 


-- 
Dr Ulf D Schiller
Centre for Computational Science
University College London
20 Gordon Street
London WC1H 0AJ
United Kingdom

Phone: +44 (0)20 7679 5300