fix

Syntax

fix group_name boolean_release boolean_def
    assign_style assign_x assign_y assign_z disp_lim
    time time_start time_end
    boolean_grad
    grad_ref_axis grad_assign_axis
    grad_ref_l grad_ref_u
  • group_name = a string (length <= 30)

  • boolean_release, boolean_def, boolean_grad = t or f

    t is true
    f is false
    
  • assign_style = disp or force

  • assign_x, assign_y, assign_z = real number or null

  • disp_lim = non-negative real number

  • time_start, time_end = non-negative integer

  • grad_ref_axis, grad_assign_axis = 1 or 2 or 3

  • grad_ref_l, grad_ref_u = real number or inf

Examples

fix big_sphere t t disp 0. null 0. 5. time 0 2500 f
fix blue_cone t t disp 5. 0. 0. 10. time 0 2500 t 2 1 50. 60.
fix top_box t t force 0. 1. 0. 3. time 0 2500 t 3 2 100. 10.

Description

This command applies displacements and/or forces to new groups and restart groups, the numbers of which are provided in the group_num command. The number of fix commands is fix_number. The new groups are created by first providing the elements/nodes/atoms information in the group command, while the same information for the restart groups, which are introduced when restart_group_number > 0, is read from group_in_*.id, where * is a positive integer starting from new_group_number + 1.

When the groups are at the simulation cell boundaries, this command is useful in applying displacement, traction, or mixed boundary conditions.

group_name must match one of the new groups or restart groups. All in this command take effect only when time_start < simulation step < time_end, unless stated otherwise.

when simulation step > time_end, the group is no longer assigned a displacement/force if boolean_release = t; the group is assigned a displacement/force vector [assign_x, assign_y, assign_z] whose non-null components are zeroed.

When boolean_def = t, the group is deformed along with the simulation box. The deformation-induced displacement is added on top of the assigned displacement/force.

assign_style = disp or force, meaning that a displacement or force vector [assign_x, assign_y, assign_z], in Angstrom/time_step or eV/Angstrom, is applied to all nodes/atoms in the group at each simulation step, after their interatomic potential-based displacement/force is discarded. If any component of the displacement/force vector is null, no displacement/force is assigned to this component. In the first example, big_sphere is fixed along the x and z directions but not along the y direction.

disp_lim is the upper bound of the magnitude of the total group displacement, in units of lattice_constant. If the total displacement magnitude (in Angstrom instead of in Angstrom/time_step) is larger than disp_lim, the displacement vector is zeroed, regardless of whether time_end is reached or what value boolean_release is. disp_lim is irrelevant when assign_style = force. However, it needs to be provided regardless.

When boolean_grad = f, the same displacement/force vector [assign_x, assign_y, assign_z] is assigned to all nodes/atoms of the group; the following options, including grad_ref_axis, grad_assign_axis, grad_ref_l, and grad_ref_u, become irrelevant and do not need to provided.

When boolean_grad = t, the displacement/force is assigned to the group gradiently, i.e., different elements/nodes/atoms in the group may have a different [assign_x, assign_y, assign_z] vector. The grad_assign_axis component of the displacement/force vector is linearly applied to the group based on the positions of elements/nodes/atoms along the grad_ref_axis direction. grad_ref_l and grad_ref_u are the bounds of the graded displacement/force, in units of the component of the lattice periodicity length vector \vec{l'}_0 along the grad_ref_axis direction, with inf referring to the lower (grad_ref_l) and upper (grad_ref_u) simulation cell boundaries.

If grad_ref_l < grad_ref_u, the elements/nodes/atoms located at or below grad_ref_l are assigned a zero displacement/force vector, i.e., fixed; those located at or above grad_ref_u are assigned [assign_x, assign_y, assign_z]. If grad_ref_l > grad_ref_u, the elements/nodes/atoms located at or above grad_ref_l are assigned a zero displacement/force vector, i.e., fixed; those located at or below grad_ref_u are assigned [assign_x, assign_y, assign_z]. In any case, the elements/nodes/atoms located between grad_ref_l and grad_ref_u are assigned a vector whose grad_assign_axis component is linearly graded while the other two components remain the same with respect to [assign_x, assign_y, assign_z].

In the second example, the elements/nodes/atoms which are located below 50.0\cdot\vec{l'}_0[2] along the y axis (because grad_ref_axis = 2) are assigned a zero displacement vector; those located above 60.0\cdot\vec{l'}_0[2] along the y axis are assigned [assign_x, assign_y, assign_z]; those in between are assigned a linearly graded displacement vector whose x component (because grad_assign_axis = 1) is varied between zero and assign_x while its y and z components are assign_y and assign_z, respectively.

In the third example, the elements/nodes/atoms which are located below 10.0\cdot\vec{l'}_0[3] along the z axis (because grad_ref_axis = 3) are assigned [assign_x, assign_y, assign_z]; those located above 100.0\cdot\vec{l'}_0[3] along the z axis are assigned a zero force vector; those in between are assigned a linearly graded force vector whose y component (because grad_assign_axis = 2) is varied between zero and assign_y while its x and z components are assign_x and assign_z, respectively.

Note that for each group concerned in this command, the displacement and force vectors are added to relevant nodes/atoms after their interatomic potential-based displacement/force vectored are zeroed. In particular, the force, stress, and energy are zeroed if assign_style = disp; the force, stress, and energy are replaced with those induced by this command if assign_style = force. In both cases, the velocity vectors are also zeroed in dynamic and hybrid CAC.

In this sense, if the same atoms/nodes are included in multiple groups that are also concerned in this command, those appearing in the later fix commands will provail. For example, if a node is assigned a displacement vector in the first fix command, a force vector in the second fix command, and another force vector in the third fix command, the force vector in the last fix command will be imposed. As another example, if an atom is assigned a force vector in the fourth fix command, and a displacement vector in the fifth fix command, the force/stress/energy vector of that atom will be zeroed. To avoid unintended effects, users are advised to carefully check if the same nodes/atoms are involved in different fix commands.

There cannot be fewer fix commands than fix_number, which should not be larger than new_group_number + restart_group_number. When there are more fix commands in cac.in than fix_number, those appearing later will be ignored.

Note that all groups do not necessarily have corresponding fix command. The purpose of having a group that does not have a correpsonding fix command is to calculate certain mechanical properties, e.g., energy, force, and stress, of the nodes/atoms it contains.

fix.f90, fix_displacement.f90, and fix_force.f90

Default

None.