DISCRETE FOURIER TRANSFORM (FFT) CONCEPTS FOR VIBRATION ANALYSIS

摘要

The FFT, the Fast Fourier transform or discrete Fourier transform, DFT, is extremely basic to our business. There is a fundamental interchangeability between time and frequency. We are able to look at a vibration signal from two equally basic points of view. There is a also duality between time and wavelet domains, and probably many others, but frequency seems most natural for us. Every chunk of signal, once digitized, can be exactly and completely saved in either the time or frequency information. There is no basic economy in data storage in either format. We use the DFT to form a spectrum of the signal sample; to form an average spectrum, to observe spectrum evolution or change as time goes on. I'll explain the time frequency transformation to make the calculations clearer, understandable. The paper also presents inverse transformation, band-wise reconstruction, and the Hilbert transform.