ABSTRACT
The importance of reliable engine diagnosis cannot be over emphasized as large
revenues are lost due to unplanned shut downs or unnecessary scheduled
maintenances. An accurate simulation model must be set up in order to minimize the
errors in performance predictions and diagnostics analysis.
The objective of the proceeding works is to have a compressor performance simulation
model matching the actual site data by a novel method and a model that can be reliably
used for performance predictions and diagnostics. In the adaptation process, Scaling
Factors (SF) are introduced and then applied to modify the compressor maps. Linear
adaptation is applied to single test point while non-linear adaptation is applied to
multiple test points. Three site real base cases are available and have been analyzed
and evaluated, at 99.6%, 93.0% and 90% RPM.
Predictions of compressor performance are compared with the manufacturer's fleet
maps. The prediction accuracy has been improved significantly after the adaptation.
Moreover, the proposed adaptation approach could be applied for most situations.