Syntax:
compute ID group-ID temp/profile xflag yflag zflag binstyle args
x arg = Nx
y arg = Ny
z arg = Nz
xy args = Nx Ny
yz args = Ny Nz
xz args = Nx Nz
xyz args = Nx Ny Nz
Nx,Ny,Nz = number of velocity bins in x,y,z dimensions
Examples:
compute myTemp flow temp/profile 1 1 1 x 10 compute myTemp flow temp/profile 0 1 1 xyz 20 20 20
Description:
Define a computation that calculates the temperature of a group of atoms, after subtracting out a spatially-averaged velocity field, before computing the kinetic energy. This can be useful for thermostatting a collection of atoms undergoing a complex flow, e.g. via a profile-unbiased thermostat (PUT) as described in (Evans). A compute of this style can be used by any command that computes a temperature, e.g. thermo_modify, fix temp/rescale, fix npt, etc.
The xflag, yflag, zflag settings determine which components of average velocity are subtracted out.
The binstyle setting and its Nx, Ny, Nz arguments determine how bins are setup to perform spatial averaging. "Bins" can be 1d slabs, 2d pencils, or 3d bricks depending on which binstyle is used. The simulation box is partitioned conceptually into Nx by Ny by Nz bins. Depending on the binstyle, you may only specify one or two of these values; the others are effectively set to 1 (no binning in that dimension). For non-orthogonal (triclinic) simulation boxes, the bins are "tilted" slabs or pencils or bricks that are parallel to the tilted faces of the box. See the region prism command for a discussion of the geometry of tilted boxes in LAMMPS.
When a temperature is computed, the velocity for the set of atoms that are both in the compute group and in the same spatial bin is summed to compute an average velocity for the bin. This bias velocity is then subtracted from the velocities of individual atoms in the bin to yield a thermal velocity for each atom. Note that if there is only one atom in the bin, it's thermal velocity will thus be 0.0.
After the spatially-averaged velocity field has been subtracted from each atom, the temperature is calculated by the formula KE = dim/2 N k T, where KE = total kinetic energy of the group of atoms (sum of 1/2 m v^2), dim = 2 or 3 = dimensionality of the simulation, N = number of atoms in the group, k = Boltzmann constant, and T = temperature.
A kinetic energy tensor, stored as a 6-element vector, is also calculated by this compute for use in the computation of a pressure tensor. The formula for the components of the tensor is the same as the above formula, except that v^2 is replaced by vx*vy for the xy component, etc. The 6 components of the vector are ordered xx, yy, zz, xy, xz, yz.
The number of atoms contributing to the temperature is assumed to be constant for the duration of the run; use the dynamic option of the compute_modify command if this is not the case.
The removal of the spatially-averaged velocity field by this fix is essentially computing the temperature after a "bias" has been removed from the velocity of the atoms. If this compute is used with a fix command that performs thermostatting then this bias will be subtracted from each atom, thermostatting of the remaining thermal velocity will be performed, and the bias will be added back in. Thermostatting fixes that work in this way include fix nvt, fix temp/rescale, fix temp/berendsen, and fix langevin.
This compute subtracts out degrees-of-freedom due to fixes that constrain molecular motion, such as fix shake and fix rigid. This means the temperature of groups of atoms that include these constraints will be computed correctly. If needed, the subtracted degrees-of-freedom can be altered using the extra option of the compute_modify command.
See this howto section of the manual for a discussion of different ways to compute temperature and perform thermostatting. Using this compute in conjunction with a thermostatting fix, as explained there, will effectively implement a profile-unbiased thermostat (PUT), as described in (Evans).
Output info:
This compute calculates a global scalar (the temperature) and a global vector of length 6 (KE tensor), which can be accessed by indices 1-6. These values can be used by any command that uses global scalar or vector values from a compute as input. See this section for an overview of LAMMPS output options.
The scalar value calculated by this compute is "intensive". The vector values are "extensive".
The scalar value will be in temperature units. The vector values will be in energy units.
Restrictions:
You should not use too large a velocity-binning grid, especially in 3d. In the current implementation, the binned velocity averages are summed across all processors, so this will be inefficient if the grid is too large, and the operation is performed every timestep, as it will be for most thermostats.
Related commands:
compute temp, compute temp/ramp, compute temp/deform, compute pressure
Default:
The option default is units = lattice.
(Evans) Evans and Morriss, Phys Rev Lett, 56, 2172-2175 (1986).