Diagnostic modules based on chaos theory for condition monitoring of rotating machinery

In various operating conditions a rotating machine and in general a nonlinear mechanical system may vibrate chaotically; this is e.g. the case of hydrodynamic instability of supports, of cracked rotors, of rotor rubbing, of gyroscopic phenomena, etc. A measure of the quantity of chaos may be used as a diagnostic parameter. The paper reports a description of various software modules based on chaos theory, developed by the authors for monitoring and diagnostics. Modules have been tested through computer simulated signals or through experimental signals. Modules take care of the fact that signals of experimental origin introduce additional problems in filtering with respect to computer simulated signals (disturbances, extremely short time histories,..). An experimental instrumented model of rotor was used to generate time histories of various instabilities. The integration of the innovating modules for monitoring and diagnostics with the traditional software and the creation of a data base for diagnosing damage and/or trend to damage will be the basis for condition monitoring.