One or more variables can be used within EnSight as a means of filtering out elements in model parts. In the simplest form a single failure variable might be specified which will remove elements that are beyond a threshold and therefore no longer viable. The element filtering is a deep filter in that elements are not just visually removed on the client, but are also removed for all calculations on the server. For example, the volume of a part calculated using the calculator will only include the non-filtered elements. If you only need to visually remove elements consider Do Element Blanking.
The filter criteria can be something as simple as a variable with two states (one state to indicate the element is failed, and the other state to indicate that the element is not failed), or it can be a variable (such as a Von Mises stress/strain) threshold for which you specify limiting values and conditions for element removal. The filter variable would ideally be a per-element variable. But you can also use a per-node variable in which case the nodal values are averaged on the element.
Note: Element filtering can only be performed on model parts. If failed elements are desired on created parts such as clip planes, isosurfaces, etc one might instead consider utilizing the Filter Part feature which will create a new part by filtering out nodes and elements from a list of parent parts.
Applying element filters to model parts will affect dependent parts such as clipping planes.
Element filtering only removes elements. It does not remove nodes. However, in most cases this limitation is accounted for such as initializing a viewport view, computing the min/max palette extents for a node based variable, etc. However, if you display nodes on the part you may very likely see nodes that are attached to elements that are blanked out or if you perform a query of the cursor tool for the closest node you will pick up one of these nodes.
Six filters are available. Each filter can specify a variable to use in the filtering process which can be compared against another variable or a constant value. The filters are combined with and or or operators. The filtering occurs sequentially through the filters, that is, it is not possible, for example, to specify a filter operation of variable1 < 0.5 OR (variable2 > 1.0 AND variable 3 < 0.0)
See Filter Parts.