Dear all,
hi,
I am confused about Poiseuille flow properties.
where should I apply the thermostat ? to wall atoms or fluid?
In the examples of Lammps, it is applied to flow atoms, but I recently noticed some papers applying it on the wall.
Thank you
Mohammad
Dear all,
hi,
I am confused about Poiseuille flow properties.
where should I apply the thermostat ? to wall atoms or fluid?
your choice. depends on what level of detail,
accuracy and convenience you need.
In the examples of Lammps, it is applied to flow atoms, but I recently
noticed some papers applying it on the wall.
the examples shipped with LAMMPS are just that: examples.
many of them are not even *meant* to show you a good way
of doing MD, but rather demonstrate how to select and use
a certain feature in the code. the examples are often simplified
to show the basic principle and they should under no
circumstances be considered "The definite guide of how use LAMMPS".
what is proper or not in a simulation often depends on
the context and thus there are no absolute rules.
you have to make your own choices and learn how to
justify them based on practical considerations as well
as the requirements of the statistical mechanical properties
of the system under consideration.
axel.
where should I apply the thermostat ? to wall atoms or fluid?
Both approaches have their advantages and drawbacks. A recurrent issue
in MD is that one has to use large forcing in order to extract the
system response out of thermal fluctuations. For liquid flows, one
need to impose huge shear rates to the system, therefore strongly
heating the liquid through viscous friction. Thermostatted walls may
not be able to remove the heat generated by the flow, possibly
resulting in temperature gradients in the liquid. It is therefore
important in this case to check that the temperature is homogeneous
inside the liquid. Thermostatting the liquid solve this problem, but
of course a thermostat affects the atom velocities, so it is not
neutral with regards to the hydrodynamic velocity. In this case it is
important to check that the thermostat does not affect the flow
significantly, e.g. by varying the coupling constant, etc. There are
several discussion in the mailing list archive on the optimal choice
of a thermostat for flow simulations.
Best,
Laurent