Abstract
Fault identification in rotor systems has been studied by many authors, but the considered malfunction is one single fault only, generally an unbalance. Real machines can be affected by several different types of faults; moreover sometimes also two different faults may develop simultaneously. A model based method for identifying multiple faults acting simultaneously on a rotor system in the frequency domain is briefly described and its robustness with regards to measuring and modelling errors is evaluated, by means of numerical simulations performed on the models of two typical power plant machines: a steam turbogenerator and a gas turbogenerator.
Highlights
In recent years the topic of the fault identification, or diagnosis, in general technical processes has been the object of a huge number of papers in different fields of engineering science
If we focus our interest on fault identification in rotating machines under a mechanical point of view, in this case several recent studies have appeared and indicated some preferred research path
A model based method for the identification of simultaneous faults in rotor systems has been presented in this paper and all the analytic details are explained
Summary
In recent years the topic of the fault identification, or diagnosis, in general technical processes has been the object of a huge number of papers in different fields of engineering science. As regards the quantitative approach, it is normally a model based fault detection method and a reliable model of the system or of the process is used for creating the symptom-fault correlation, or the input-output relation. This method has many different ways of applications, someone of them merge qualitative and quantitative approaches. A systematic approach has been introduced by the authors in [9] to identify several different types of faults and to discriminate among faults with similar harmonic components This method has been experimentally validated on different test-rigs and real machines (see [10,11,12,13]) for many types of faults, such as unbalances, rotor permanent bows, rotor rubs, coupling misalignments, cracks, journal ovalization and rotor stiffness asymmetries. The identification of the multiple faults is made by a least square fitting in the frequency domain, by means of the minimization of a multidimensional residue between the measured vibration in some measuring planes on the machine (usually, but not necessarily, the bearings) and the calculated vibrations due to the acting faults
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
