Application of signal analysis for diagnostics

Signals are usually described in two domains: time and frequency. The Fourier transform is used for signal transformation from time domain to frequency domain and vice versa and that is enough to analyse the stationary signal. If a signal is non-stationary, the Fourier transform can be utilized to determine frequencies in signal but it can't be applied to determine the moment when a particular signal exists. For the analysis of non-stationary signals the wavelet transform can be used as well as a continuous or discrete wavelet transform. If the signal has a large number of samples, a discrete wavelet transform should be applicable preferably. If the signal is a discrete dynamic series, fast algorithms of discrete wavelet transforms can be used. The article presents 4 types of wavelets that can be applied to carry out fast discrete wavelet transforms. Based on the performed calculations the most appropriate wavelet is chosen.