Use of modal representation for the supporting structure in model-based fault identification of large rotating machinery: Part 2—application to a real machine

Model-based techniques are often employed in diagnostics of rotating machines to locate and to evaluate the severity of a malfunction. The use of a reliable model can increase the accuracy of identification. Rigid supports or lumped mass pedestals are not always enough to account for foundation dynamics; a modal representation of the supports can improve the identification results. The method, discussed in the first part, is here validated using experimental data of a 320 MW steam turbogenerator. To the authors’ knowledge, this is the first case of fault identification on real data from a large machine, where the supporting structure is accounted for by means of a modal model.