cochiece
2008-03-25, 11:57
Hi all,
I'm currently trying to write an explicit subroutine for an 8-node brick user element (VUEL). To validate it I've defined a single element with an ENCASTRE constraint on one face and a velocity boundary condition imposed on the opposite face. I've defined the mass matrix, stiffness matrix (k) and stable time increment but defining the load vector (f) is confusing me...
Abaqus provides the nodal displacements (u) at the start of each increment and so I thought I could simply define the load as f = k*u (there's no damping or acceleration in the model at the moment). Unfortunately once Abaqus returns f to the model it causes certain nodel displacement to become very large. In the next iteration then f is over estimated due to these large nodal displacements and both f and u cascade out of control until the element explodes.
Has anyone here encountered similar problems in user element or even experience in general with them?
Thanks!
I'm currently trying to write an explicit subroutine for an 8-node brick user element (VUEL). To validate it I've defined a single element with an ENCASTRE constraint on one face and a velocity boundary condition imposed on the opposite face. I've defined the mass matrix, stiffness matrix (k) and stable time increment but defining the load vector (f) is confusing me...
Abaqus provides the nodal displacements (u) at the start of each increment and so I thought I could simply define the load as f = k*u (there's no damping or acceleration in the model at the moment). Unfortunately once Abaqus returns f to the model it causes certain nodel displacement to become very large. In the next iteration then f is over estimated due to these large nodal displacements and both f and u cascade out of control until the element explodes.
Has anyone here encountered similar problems in user element or even experience in general with them?
Thanks!