10 docs tagged with "constraints"

One further constraint that users should be aware of is the Automatic Constraint. These are constraints set up by the Gss solver to enable the solution of shell models where the elements do not have stiffness in all directions. Shell elements do not have a local zz degree of freedom. Thus when the elements are transformed into a non-global plane they appear to have stiffness is all directions but one of the principal stiffnesses at the nodes is zero. To avoid problems of zero stiffness at the solution stage these degrees of freedom have to be removed from the solution. The method of doing this is termed applying Automatic Constraints.

There are various rules that must be followed when specifying constraints. Failure to abide by these rules results in constraints that are in conflict. Such conflicts are reported by the pre-analysis data checking.

The constraint axis of a node (also referred to as the local axis of a node) is the axes that determines the directions of constraints (such as restraints, joints, rigid constraints, constraint equations) or local axis loads. Spring, mass and damper properties on a node also act in the constraint axis of the node.

Constraint equations allow a node, in a particular direction to be constrained relative to a set of other nodes. Constraint equations are the fundamental building block for all other constraint types.

In the most general sense constraints are where the user is putting some restriction on the free displacement of the structure. Some constraint is required on all structures to prevent pure rigid body motions. Thus in general the minimum number of constraints on a 3D structure is six, three suppressing the translational rigid body modes and three suppressing the rotational rigid body modes.

Rigid constraints can be created using the Sculpt > Constraint operations > Create rigid constraint menu command. The procedure is as follows.

Joints allow nodes to be tied in the specified translational or rotational directions. Unlike rigid constraints, joints do not impose equilibrium on the model.

At times there are very stiff sections in a model and generally it is not good to model these simply by creating very stiff elements as this can lead to ill-conditioning problems. It is better to replace the elements in these regions with link elements or rigid constraints. The action of link elements and rigid constraints is substantially the same, the main difference being that a rigid constraint can include many nodes while a link element can only include two nodes, but has some additional linkage options.

Rigid constraints define sets of nodes constrained to move as a rigid body. These are commonly use to model rigid diaphragms, where typically the nodes are rigid in the x-y plane but not in the z direction.

Tied interfaces allow two sets of elements that do not have connectivity to be joined without the need for complex mesh refinement. Instead the elements on one side are taken as primary and the other side as constrained. The nodes on the boundary of the constrained side are tied to the nodes on the boundary of the primary side through a set of constraint equations.