STUDY AND PERFORMANCE EVALUATION OF SOME NONLINEAR DIAGNOSTIC METHODSFOR LARGE ROTATING MACHINERY

In this paper, some nonlinear diagnostic methods for large rotating ma chinery are introduced and evaluated from the practitioners' points of view, i.e. their sensitivity to nonlinear phenomena, their applicabil ity in machinery diagnostic practice, their effectiveness in computer execution and finally, their acceptability in the enterprises. The eva luated methods are pseudo-phase diagrams or limit cycle detection, sin gular spectrum analysis (SSA), the Wiper distribution and the Kullback index of the complexity based on information theory. Examples show th e pseudo-phase diagrams are the most convenient for on-line diagnostic s. It can be directly applied in the productional practice, while all the other methods can be used as supplementary diagnostic measures.