Choosing the Reference Model

Plot relative to the current model

This is the default case. You select a reference state in the current model, and results are plotted relative to this state. If you have multiple models in a window then each model will be plotted relative to itself.

Data from the reference state are subtracted from those in the current state, and the results displayed. It is perfectly possible to choose a reference state later than the current one, and hence to get "negative" results, the computation is simply:

<data> _displayed = <data> _current - <data> _reference

Plot relative to another model

This is the more complex case of plotting data from this model relative to a state from a different model. If you have multiple models in a window each model in that window will be plotted relative to the reference model.

The principle is exactly the same: the reference data is subtracted from the current, but mapping of reference model onto current is done as follows:

• Mapping is "by external label".
  
  The results from node label <i> in the reference are subtracted from those for node label <i> in the current model. And likewise for elements.

• No checking for geometrical or topological proximity takes place.
  
  No check is made that node <i> (or element <j>) in the two models are equivalent, or even remotely in the same place - either topologically or geometrically.

• If no equivalent label is found, zero is reported.
  
  If no matching node or element can be found in the reference model, then zero is reported as a result - regardless of the actual value in the current model.

This means that models which are topologically nearly identical compare well, but areas which have been remeshed may give very misleading comparisons.