Vibration-Based Condition Monitoring in Rotating Machinery

In Chapters 5 and 6 , measurement and signal processing techniques, transducers, signal conditioners, and signal analysis equipment, which are used in rotating machinery condition monitoring and identification, are described. It is vital to display the measured signal in suitable form so as to be beneficial for the elucidation of the disorder of rotating machinery. In this chapter, by looking at numerous forms of measured signals, the likely condition of the machinery is provided. It also looks into the correlation of a particular signature with a particular failure in more detail. Every fault has a particular signature in the measured signal, and it is the most suitable and inexpensive technique to isolate a likely fault in machinery. Now very advanced signal processing and machine learning techniques [fast Fourier transform (FFT), wavelet transform, full spectrum, neural network, fuzzy logic, genetic algorithms (GAs), genetic programming, and support vector machine (SVM)] are being applied to the vibration signals of laboratory test setups and of actual rotating machinery to detect, locate, and quantify the faults, and based on this, the life of the machinery is also being predicted. A brief review of the applications of these techniques for rotating machinery condition monitoring is also provided since detailed treatment of these newly emerging methods is beyond the scope of this book. Proactive action to prevent a failure is better than the detection of failure. Chapter 8 will deal with introduction of the active control of rotors by magnetic bearings, which is still an area of research, and its application in the field of rotor dynamics, especially in the condition monitoring and the system identification.