转发：回复： doubt about fix addforce which is time-dependent

LAMMPS LAMMPS Mailing List Mirror
1

Hi all:
I am trying to simulate the fracture of cuboid rotating blade by
peridynamics.I came up with two problems.
(1) I use "fix move rotate"command to rotate the blade around the x-axis.
My doubt is whether it will produce centrifugal force automatically without
imposinging any other force?

think a little bit about where the centrifugal force is coming from.
best you consider a minimal system with just one fixed atom and a
second connected to it with a harmonic potential. you make the second
rotate around the first by giving it a push.

fix move will just move atoms on a prescribed trajectory without
paying any attention to forces. where would the centrifugal force
come from

(2) I want to impose a tension force to the rotating blade to simulate
the fracture.The blade rotates around the x-axis.The force is a constant
along the longitudinal direction of the blade,and is vertical to x-axis
always.So i use “fix addfore” command to define this time-dependent force
without using the “minimize” command.
But i think this time-dependent force is not recognized likely, why?

worse. your script is plain wrong. you use fix move to propagate
all(!) particles and then you add fix nve to do the same. haven’t you
seen the warning that LAMMPS throws in this case?

you have two options:

  • you impose the rotation like with a real blade at the axle only and
    let the rest of the system follow. case closed. applying an external
    force in addition would make no sense.
  • you treat the centrifugal force as an external force (and make the
    rotating blade your reference system). in that case you don’t need the
    rotation at all. apply the external force and integrate the equations
    of motion. done.

these are all problems of rather elementary classical mechanics. it
may be helpful to dig around for some of your very old text books that
cover those topics and re-read them.

axel.

2

发件人: "松行天下"<[email protected]...>;
发送时间: 2013年6月3日(星期一) 下午3:02
收件人: "Axel Kohlmeyer"<[email protected]>;
主题: 回复： [lammps-users] doubt about fix addforce which is time-dependent

Dear Axel Kohlmeyer:
    Thanks for your reply！
    I want to research the trajectory and energy of blade debris after the
rotating blade broken.So,I think ,firstly, the blade must rotate.

no. not really. remember that you are simulating a system in vacuum,
so it doesn't make a difference whether you have a rotating reference
frame or a fixed frame.

Secondly, i should impose a force to the blade to achieve fracture.

as i wrote you _might_ impose that through the rotation, if you impose
the rotation only to a small subset of particles and also properly
initialize the velocities of all particles. in that case no external
force is needed.

but if you prefer to impose an external force gradient instead, then
you don't need the rotation and the problem is much easier simulated
in a fixed reference frame.

    Now ,the blade rotation has been achieved, but i do not know how to
impose a force to the rotating blade? The force is a constant along the

have you fixed your input? it was grossly wrong. also, you should
*really* spend a little more time considering what your system is
like.

axel.