Files
lammps/doc/src/pair_dpd_fdt.rst
2024-08-15 22:52:42 -04:00

187 lines
6.8 KiB
ReStructuredText

.. index:: pair_style dpd/fdt
.. index:: pair_style dpd/fdt/energy
.. index:: pair_style dpd/fdt/energy/kk
pair_style dpd/fdt command
==========================
pair_style dpd/fdt/energy command
=================================
Accelerator Variants: *dpd/fdt/energy/kk*
Syntax
""""""
.. code-block:: LAMMPS
pair_style style args
* style = *dpd/fdt* or *dpd/fdt/energy*
* args = list of arguments for a particular style
.. parsed-literal::
*dpd/fdt* args = T cutoff seed
T = temperature (temperature units)
cutoff = global cutoff for DPD interactions (distance units)
seed = random # seed (positive integer)
*dpd/fdt/energy* args = cutoff seed
cutoff = global cutoff for DPD interactions (distance units)
seed = random # seed (positive integer)
Examples
""""""""
.. code-block:: LAMMPS
pair_style dpd/fdt 300.0 2.5 34387
pair_coeff * * 3.0 1.0 2.5
pair_style dpd/fdt/energy 2.5 34387
pair_coeff * * 3.0 1.0 0.1 2.5
Description
"""""""""""
Styles *dpd/fdt* and *dpd/fdt/energy* compute the force for dissipative
particle dynamics (DPD) simulations. The *dpd/fdt* style is used to
perform DPD simulations under isothermal and isobaric conditions,
while the *dpd/fdt/energy* style is used to perform DPD simulations
under isoenergetic and isoenthalpic conditions (see :ref:`(Lisal) <Lisal3>`).
For DPD simulations in general, the force on atom I due to atom J is
given as a sum of 3 terms
.. math::
\vec{f} = & (F^C + F^D + F^R) \hat{r_{ij}} \qquad \qquad r < r_c \\
F^C = & A w(r) \\
F^D = & - \gamma w^2(r) (\hat{r_{ij}} \bullet \vec{v}_{ij}) \\
F^R = & \sigma w(r) \alpha (\Delta t)^{-1/2} \\
w(r) = & 1 - \frac{r}{r_c}
where :math:`F^C` is a conservative force, :math:`F^D` is a dissipative
force, and :math:`F^R` is a random force. :math:`\hat{r_{ij}}` is a
unit vector in the direction :math:`r_i - r_j`, :math:`\vec{v}_{ij}` is
the vector difference in velocities of the two atoms, :math:`\vec{v}_i -
\vec{v}_j`, :math:`\alpha` is a Gaussian random number with zero mean
and unit variance, *dt* is the timestep size, and :math:`w(r)` is a
weighting factor that varies between 0 and 1, :math:`r_c` is the
pairwise cutoff. Note that alternative definitions of the weighting
function exist, but would have to be implemented as a separate pair
style command.
For style *dpd/fdt*, the fluctuation-dissipation theorem defines
:math:`\gamma` to be set equal to :math:`\sigma^2/(2 T)`, where *T* is the
set point temperature specified as a pair style parameter in the above
examples. The following coefficients must be defined for each pair of
atoms types via the :doc:`pair_coeff <pair_coeff>` command as in the
examples above, or in the data file or restart files read by the
:doc:`read_data <read_data>` or :doc:`read_restart <read_restart>`
commands:
* A (force units)
* :math:`\sigma` (force\*time\^(1/2) units)
* cutoff (distance units)
The last coefficient is optional. If not specified, the global DPD
cutoff is used.
Style *dpd/fdt/energy* is used to perform DPD simulations under
isoenergetic and isoenthalpic conditions. The fluctuation-dissipation
theorem defines :math:`\gamma` to be set equal to :math:`\sigma^2/(2
\theta)`, where :math:`\theta` is the average internal temperature for
the pair. The particle internal temperature is related to the particle
internal energy through a mesoparticle equation of state (see :doc:`fix
eos <fix>`). The differential internal conductive and mechanical
energies are computed within style *dpd/fdt/energy* as:
.. math::
du_{i}^{cond} = & \kappa_{ij}(\frac{1}{\theta_{i}}-\frac{1}{\theta_{j}})\omega_{ij}^{2} + \alpha_{ij}\omega_{ij}\zeta_{ij}^{q}(\Delta{t})^{-1/2} \\
du_{i}^{mech} = & -\frac{1}{2}\gamma_{ij}\omega_{ij}^{2}(\frac{\vec{r}_{ij}}{r_{ij}}\bullet\vec{v}_{ij})^{2} -
\frac{\sigma^{2}_{ij}}{4}(\frac{1}{m_{i}}+\frac{1}{m_{j}})\omega_{ij}^{2} -
\frac{1}{2}\sigma_{ij}\omega_{ij}(\frac{\vec{r}_{ij}}{r_{ij}}\bullet\vec{v}_{ij})\zeta_{ij}(\Delta{t})^{-1/2}
where
.. math::
\alpha_{ij}^{2} = & 2k_{B}\kappa_{ij} \\
\sigma^{2}_{ij} = & 2\gamma_{ij}k_{B}\Theta_{ij} \\
\Theta_{ij}^{-1} = & \frac{1}{2}(\frac{1}{\theta_{i}}+\frac{1}{\theta_{j}})
:math:`\zeta_ij^q` is a second Gaussian random number with zero mean and
unit variance that is used to compute the internal conductive
energy. The fluctuation-dissipation theorem defines :math:`alpha^2` to
be set equal to :math:`2k_B\kappa`, where :math:`\kappa` is the
mesoparticle thermal conductivity parameter. The following coefficients
must be defined for each pair of atoms types via the :doc:`pair_coeff
<pair_coeff>` command as in the examples above, or in the data file or
restart files read by the :doc:`read_data <read_data>` or
:doc:`read_restart <read_restart>` commands:
* A (force units)
* :math:`\sigma` (force\*time\^(1/2) units)
* :math:`\kappa` (energy\*temperature/time units)
* cutoff (distance units)
The last coefficient is optional. If not specified, the global DPD
cutoff is used.
The pairwise energy associated with styles *dpd/fdt* and
*dpd/fdt/energy* is only due to the conservative force term :math:`F^C`,
and is shifted to be zero at the cutoff distance :math:`r_c`. The
pairwise virial is calculated using only the conservative term.
The forces computed through the *dpd/fdt* and *dpd/fdt/energy* styles
can be integrated with the velocity-Verlet integration scheme or the
Shardlow splitting integration scheme described by :ref:`(Lisal)
<Lisal3>`. In the cases when these pair styles are combined with the
:doc:`fix shardlow <fix_shardlow>`, these pair styles differ from the
other dpd styles in that the dissipative and random forces are split
from the force calculation and are not computed within the pair style.
Thus, only the conservative force is computed by the pair style, while
the stochastic integration of the dissipative and random forces are
handled through the Shardlow splitting algorithm approach. The Shardlow
splitting algorithm is advantageous, especially when performing DPD
under isoenergetic conditions, as it allows significantly larger
timesteps to be taken.
----------
.. include:: accel_styles.rst
----------
Restrictions
""""""""""""
These commands are part of the DPD-REACT package. They are only enabled
if LAMMPS was built with that package. See the :doc:`Build package
<Build_package>` page for more info.
Pair styles *dpd/fdt* and *dpd/fdt/energy* require use of the
:doc:`comm_modify vel yes <comm_modify>` option so that velocities are
stored by ghost atoms.
Pair style *dpd/fdt/energy* requires :doc:`atom_style dpd <atom_style>`
to be used in order to properly account for the particle internal
energies and temperatures.
Related commands
""""""""""""""""
:doc:`pair_coeff <pair_coeff>`, :doc:`fix shardlow <fix_shardlow>`
Default
"""""""
none
----------
.. _Lisal3:
**(Lisal)** M. Lisal, J.K. Brennan, J. Bonet Avalos, J. Chem. Phys., 135, 204105 (2011).