Zero length elements definition - eesd-epfl/OpenSees GitHub Wiki
The definition of an equivalent frame model through this macroelement requires in particular the explicit modelling of wall-to-wall connections and floor-to-wall connections, as well as the floor behaviour.
Wall-to-wall and wall-to-floor connections can be modelled by zero-length elements. If the response in different directions is not coupled, as for wall-to-wall connections, standard zero-length elements can be used:
# standard zero length element with no coupling (one line)
element zeroLength $eTag $nI $nJ -mat $mat1 $mat2 ... -dir $dir1 $dir2 ... -orient $x1 $x2 $x3 $y1 $y2 $y3
Where:
| eTag | element tag |
| nI | first node of the element |
| nJ | second node of the element |
| mat1, mat2 | list of uniaxial material models to apply to the different directions |
| dir1, dir2 | list of directions to which the material models are applied |
| x1, x2, x3 | direction vector of the local x axis |
| y1, y2, y3 | vector laying in the local xy plane |
If there is the need of considering the coupling of the material response in different directions (for example for frictional models) ZeroLengthND elements can be used, to which a nDMaterial can be attached:
# ND zero length element with coupling (one line)
element zeroLengthND $eTag $nI $nJ $matND -orient $x1 $x2 $x3 $y1 $y2 $y3
| eTag | element tag |
| nI | first node of the element |
| nJ | second node of the element |
| matND | ND material model to be applied |
| x1, x2, x3 | direction vector of the local x axis |
| y1, y2, y3 | vector laying in the local xy plane |
If the orientation is not specified the local axis have the same orientation as global axes. To define the orientation of the interface in the right sense, one must consider that a displacement in the (positive) x local direction of node J, keeping node I in its position, has to correspond to an opening of the interface.
Since zero-length element need to be defined between nodes that are truly in the same location, often there is the need of duplicating nodes in the same location (for floor-to-wall interfaces) or creating new nodes connected to their master node by a rigid link. The latter procedure does not increase the number of degrees of freedom of the problem to solve, as the degrees of freedom of the new node are considered linearly dependent on the degrees of freedom of the master node and do not constitute additional variables to solve for.
To exploit this advantage, the rigid link must not be created using a stiff beam element, but a rigid link can be imposed between the master and the slave node as a constraint, with the following syntax.
rigidLink beam $masterNode $slaveNode
constraints Transformation
The constraint handler, to be specified just once at the moment of initialisation of the system, has to be of the kind “Transformation” to be able to handle linear relations between degrees of freedom expressed in matrix form, such as the ones of a rigid beam. It is important to remind that the internal node of the macroelement is equipped with two sets of translational degrees of freedom, therefore it is not suitable to be used in combination with rigid links of beam type (but can be connected through equalDOF links, if this makes sense).