More precisely, we use data reconciliation systems for obtaining
the most probable value for process data acquired on-line.
However, these data reconciliation systems do not work
properly when aberrant values (such as zeros, max values,
oscillating values) are present.
Therefore we are looking for technologies able to
identify/diagnose the presence of these aberrant values
before applying the data reconciliation software.
Any pointer would be helpful, thank you.
Bertrand Braunschweig
Institut Francais du Petrole
------------------------------